The Agrarian Studies Series at Yale University Press seeks to publish

The Art of
outstanding and original interdisciplinary work on agriculture and rural
society - for any period, in any location. Works of daring that question existing
paradigms and fill abstract categories with the lived-experience of rural people
Not Being Governed
are especially encouraged.
An Anarchist History of Upland Southeast Asia
- JAMES C. SCOTT, Series Editor

James C. Scott

Yale University Press NEW HAVEN & LONDON

 CHAPTER 1

Hills, Valleys, and States

An Introduction to Zomia

I
open with three diagnostic expressions of frustration. The first two are
from would-be conquering administrators, determined to subdue a re-
calcitrant landscape and its fugitive, resistant inhabitants. The third,
from a different continent, is from a would~be conqueror of souls, in
some despair at the irreligion and heterodoxy that the landscape appears to
encourage:
Making maps is hard, but mapping Guizhou province especially so The
land in southern Guizhou has fragmented and confused boundaries A de-
partment or a county may be split into several subsections, in many instances
separated by other departments or counties. . . . There are also regions of no
man's land where the Miao live intermixed with the Chinese....
Southern Guizhou has a multitude ofmountain peaks. They are jumbled
together, without any plains or marshes to space them out, or rivers or water
courses to put limits to them., They are vexingly numerous and ill-disciplined.
. . . Very few people dwell among them, and generally the peaks do not have
names. Their configurations are difficult to discern clearly, ridges and summits
seeming to be the same. Those who give an account of the arterial pattern, of
the mountains are thus obliged to speak at length. In some cases, to descdbe a
few kilometers of ramifications needs a pile of documentation, and dealing with
the main line of a day's march takes a sequence of chapters.
As to the confusion of the local patois, in the space of fifty kilometers a
river may have fifty names and an encampment covering a kilometer and a half
may have three designations. Such is the unreliability of the nomenclature.l
2 HILLS, VALLEYS, AND STATES
The hilly and jungly tracts were those in which the dacoits held out long-
est. Such were [sz'c] the country between Minbu and Thayetmyo and the terai
[swampy lowland belt] at the foot of the Shan Hills and the Arakan and Chin
Hills. Here pursuit was impossible. The tracts are narrow and tortuous and ad-
mirably suited for ambuscades. Except by the regular paths there were hardly
any means of approach; the jungle malaria was fatal to our troops; a column
could only penetrate the jungle and move on. The villages are small and far
between; they are generally compact and surrounded by dense, impenetrable
jungle. The paths were either just broad enough for a cart, or very narrow, and,
where they led through the jungle were overhung with brambles and thorny
creepers. A good deal of the dry grass is burned in March, but as soon as the
rains recommence the whole once more becomes impassible.2
The surface has been minutely trenched by winding streams. So numerous are
the creeks that the topographical map of a single representative county of 373
square miles indicated 339 named streams, that is, nine streams for each ten
square miles. The valleys are for the most part "V"-shaped, with rarely more
level space along the banks of a stream for a cabin and perhaps a garden patch.
. . . The isolation occasioned by methods of travel so slow and difficult is in-
tensified by several circumstances. For one thing, the routes are round-about.
Travel is either down one branch along a creek and up another branch, or up
a stream to a divide and down another stream on the further side of the ridge.
This being the case, married women living within ten miles of their parents
have passed a dozen years without going back to see them. 3
Behind each lament lies a particular project of rule: Han rule under
the Q!ng, British rule within the Empire, and finally, the rule of orthodox
Protestant Christianity in Appalachia. All would style themselves, unself-
consciously, as bearers of order, progress, enlightenment, and civilization. All
wished to extend the advantages of administrative discipline, associated with
the state or organized religion, to areas previously ungoverned.
How might we best understand the fraught dialectical relations between
such projects of rule and their agents, on the one hand, and zones of relativ~
autonomy and their inhabitants, on the other? This relationship is particu-
larly salient in mainland Southeast Asia, where it demarcates the greatest
social cleavage that shapes much of the region's history: that between hill
peoples and valley peoples or between upstream (hulu in the Malay world)
and downstream (hilir) peoples.4 In tracing this dialectic with some care, I
believe it also traces a path to a novel historical understanding of the global
process of state formation in the valleys and the peopling of the hills.
I "
HILLS, VALLEYS, AND STAT'ES 3
The encounter between expansionary states and self-governing peoples
is hardly confined to Southeast Asia. It is echoed in the cultural and admin-
istrative process of "internal colonialism" that characterizes the formation of
most modern Western, nation;..states; in the imperial projects of the Romans,
the Hapsburgs, the Ottomans, the Han, and the British; in the subjugation
of indigenous peoples in "white-settler" colonies such as the United States,
Canada, South Africa, Australia, and Algeria; in the dialectic between seden-
tary, town-dwelling Arabs<and nomadic pastoralists that have characterized
much of Middle Easterrt history. 5 The precise shape of the encounters is,
to be sure, unique to each case. Nevertheless, the ubiquity of the encounter
between self-governing and state-governed peoples-variously styled as the
raw and the cooked, the wild and the tamed, the hill/forest people and
the valley/cleared-land people, upstream and downstream, the barbarian and
the civilized, the backward and the modern, the free and the bound, the
people without history ,and the people with history - provides us with many
possibilities for comparative triangulation. We shall take advantage of these
opportunities where we can.
A World of Peripheries
In the written record - that is to say, from the beginning of grain-based,
agrarian civilizations-the encounter we are examining can fairly be said to
preoccupy rulers. But if we stand back and widen the historical lens still
further, seeing the encounter in human rather than state-civilization terms,
it is astonishing how recent and rapid the encounter has been. Homo sapiens
sapiens has been around for something like two hundred thousand years, and
only about sixty thousand, at the outside, in Southeast Asia. There the re-
gion's first small concentrations of sedentary populations appear not earlier
than the first millennium before the common era (CE) and represent a mere
smudge in the historical landscape -localized, tenuous, and evanescent.
Until shortly before the common era, the very last I percent of human his-
tory, the social landscape consisted of elementary, self-governing, kinship
units that might, occasionally, cooperate in hunting, feasting, skirmishing,
trading, and peacemaking. It did not contain anytp.ing one could call a state. 6
In other words, living in the absence of state structures has been the standard
human condition.
The founding of agrarian states, then, was the contingent event that
created a distinction, hence a diaiectic, between a settled, state-governed
 4 HILLS, VALLEYS, AND STATES
population and a frontier penumbra of less governed or virtually autono-
mous peoples. Until at least the early nineteenth century, the difficulties of
transportation, the state of military technology, and, above all, demographic
realities placed sharp limits on the reach of even the most ambitious states.'
Operating in a population density of only 5.5 persons per square kilometer
in r600 (compared with roughly 35 for India and China), a ruler's subjects in
Southeast Asia had relatively easy access to a vast, land-rich frontier? That
frontier operated as a rough and ready homeostatic device; the more a state
pressed its subjects, the fewer subjects it had. The frontier underwrote popu-
lar freedom. Richard O'Connor captures this dialectic nicely: "Once states
appeared, adaptive conditions changed yet again-at least for farmers. At
that moment, mobility allowed farmers to escape the impositions of states
and their wars. I call this tertiary dispersion. The other two revolutions-
agriculture and complex society-were secure but the state's domination of.
its peasantry was not, and so we find a strategy of 'collecting people ... and
establishing villages.'" 8
The Last Enclosure
Only the modern state, in both its colonial and its independent guises, has
had the resources to realize a project of rule that was a mere glint in the eye
of its precolonial ancestor: namely to bring nonstate spaces and people to
. heel. This project in its broadest sense represents the last great enclosure
movement in Southeast Asia. It has been pursued-albeit clumsily and with
setbacks-consistently for at least the past century. Governments, whether
colonial or independent, communist or neoliberal, populist or authoritarian,
have embraced it fully. The headlong pursuit ofthis end by regimes otherwise
starkly different suggests that such projects of administrative, economic, and
cultural standardization are hard-wired into the architecture of the modern
state itself.
Seen from the stat~ center, this enclosure movement is, in part, an effort
to integrate and monetize the people, lands, and resources of the periphery so
that they become, to use the French term, rentable-auditable contributors
to the gross national product and to foreign exchange. In truth, peripheral
peoples had always been firmly linked economically to the lowlands and to
world trade. In some cases, they appear to have provided most ofthe products
valued in international commerce. Nevertheless, the attempt to fully incor-
porate them has been culturally styled as development, economic progress,
HILLS, VALLEYS, AND STATES 5
literacy, and social integration. In practice, it has meant something else. The
objective has been less to,.make them productive than to ensure that their
economic activity was legible, taxable, assessable, and confiscatable or, fail-
ing that, to replace it with forms of production that were. Everywhere they
could, states have obliged mobile, swidden cultivators to settle in permanent
villages. They have tried to replace open common-property land tenure with
closed common property: collective farms or, more especially, the'individual
freehold property of libtrrat'(economies. They have seized timber and mineral
resources for the national patrimony. They have encouraged, whenever pos-
sible, cash, monocropping, plantation-style agriculture in place of the more
! ,:
biodiverse forms of cultiv~tion that prevailed earlier. The term enclosure
seems entirely appropriate for this process, mimicking as it does the English
enclosures that, in the century after r76r, swallowed half of England's com-
mon arable land in favor of large-scale, private, commercial prodaction.
The novel and revolutionary aspect of this great enclosure movement
is apparent if we open our historical lens to its widest aperture. The very
earliest states in China and Egypt-and later, Chandra-Gupta India" classical
Greece, and republican Rome-were, in demographic terms, insignificant.
They occupied a minuscule portion of the world's landscape, and their sub-
jects were no more than a rounding error in the world's population figures.
In mainland Southeast Asia, where the first states appear only around the
middle of the first millennium of the common era, their mark on the land-
scape and its peoples is relatively trivial when compared with their over-
sized place in the history books. Small, moated, and walled centers together
with their tributary villages, these little nodes of hierarchy and power were
both unstable and geographically confined. To an eye not yet hypnotized by
archeological remains and state-centric histories, the landscape would have
seemed virtually all periphery and no centers. Nearly all the population and
territory were outside their ambit.
Diminutive though these state centers were, they possessed a singular
strategic and military advantage in their capacity to concentrate manpower
and foodstuffs in one place. Irrigated rice agriculture on permanent fields
was the key.9 As a new political form, the padi state was an ingathering of
previously stateless peoples. Some subjects were no doubt attracted to the
possibilities for trade, wealth, and status available at the court centers, while
others, almost certainly the. majority, were 'Captives and slaves seized in war-'
fare or purchased from slave-raiders. The vast "barbarian" periphery of these
small states was a vital resource in at least two respects. First, it was the
6 HILLS, VALLEYS, AND STATES
source of hundreds of important trade goods and forest products necessary
to the prosperity of the padi state. And second, it was the source of the most
important trade good in circulation: the human captives who formed the
working capital of any successful state. What we know of the classical states
such as Egypt, Greece, and Rome, as well as the early Khmer, Thai, and Bur-
mese states, suggests that most of their subjects were formally unfree: slaves,
captives, and their descendants.
The enormous ungoverned periphery surrounding these minute states
also represented a challenge and a threat. It was home to fugitive, mobile
populations whose modes of subsistence - foraging, hunting, shifting cul-
tivation, fishing, and pastoralism - were fundamentally intractable to state
appropriation. The very diversity, fluidity, and mobility of their livelihoods
meant that for an agrarian state adapted to sedentary agriculture, this ungov-
erned landscape and its people were fiscally sterile. Unless they wished to
trade, their production was inaccessible for yet another re~son. Whereas the
early states were nearly everywhere the creature ofarable plains and plateaus,
much of the more numerous ungoverned population lived, from a state per-
spective, in geographically difficult terrain: mountains, marshland, swamps,
arid steppes, and deserts. Even if, as was rarely the case, their products were
in principle appropriable, they were effectively out of range owing to disper-
sal and the difficulties of transportation. The two zones were ecologically
complementary and therefore natural trading partners, but such trade could
rarely be coerced; it took the form of voluntary exchange.
1 1;
For early state elites, the periphery - seen frequently as the realm of
"barbarian tribes" - was also a potential threat. Rarely - but memorably, in
the case of the Mongols and the Huns and Osman and his conquering band-
a militarized pastoral people might overrun the state and destroy it or rule in
its place. More commonly, nonstate peoples found it convenient to raid the
settlements of sedentary farming communities subject to the state, some-
times exacting systematic tribute from them in the manner of states. Just as
states encouraged sedentary agriculture for its "easy pickings," so, too, did
raiders find it attractive as a site of appropriation.
The main, long-run threat of the ungoverned periphery, however, was
that it represented a constant temptation, a constant alternative to life within
the state. Founders of a new state often seized arable land from its previous
occupants, who might then either be incorporated' or choose to move away.
Those who fled became, one might say, the first refugees from state power,
joining others outside the state's reach. When and if the state's reach ex-
panded, still others faced the same dilemma.
HILLS, VALLEYS, AND STATES 7
At a time when the state seems pervasive and inescapable, it is easy to
forget that for much of history, living within or outside the state-or in an
intermediate zone-was a choice, one that might be revised as the circum-
stances warranted. A wealthy and peaceful state center might attract a grow-
ing population that found its advantages rewarding. This, of course, fits the
standard civilizational narrative of rude barbarians mesmerized by the pros-
perity made possible by the king's peace and justice-a narrative shared by
most of the world's salvational religions, not to mention Thomas Hobbes.
This narrative ignores two capital facts. First, as we have noted, it ap-
pears that much, if not most, of the population of the early states was unfree;
they were subjects under duress. The second fact, most inconvenient for the
standard narrative of civilization, is that it was very common for state sub-
jects to run away. Living within the state meant, virtually by definition, taxes,
conscription, corvee labor, and, for most, a condition of servitude; these con-
ditions were at the core of the state's strategic and military advantages. When
these burdens became overwhelming, subjects moved with alacrity to the
periphery or to another state. Under premodern conditions, the crowding of
population, domesticated animals, and the heavy reliance on a single grain
had consequences for both human and crop health that made famines and
epidemics more likely. And finally, the early states were warmaking machines
as well, producing hemorrhages of subjects fleeing conscription, invasion,
and plunder. Thus the early state extruded populations as readily as it ab-
sorbed them, and when, as was often the case, it collapsed altogether as the
result of war, drought, epidemic, or civil strife over succession, its popula-
tions were disgorged. States were, by no means, a once-and-for-all creation.
Innumerable archeologiCal finds of state centers that briefly flourished and
were then eclipsed by warfare, epidemics, famine, or ecological collapse de-
pict a long history of state formation and collapse rather than permanence.
For long periods people moved in and out of states, and "stateness" was,
itself, often cyclical and reversible. lO
This pattern of state-making and state-unmaking produced, over time,
a periphery that was composed as much of refugees as of peoples who had
never been state subjects. Much of the periphery of states became a zone of
refuge or "shatter zone," where the human shards of state formation and
rivalry accumulated willy nilly, creating regions of bewildering ethnic and
linguistic complexity. State expansion and collapse often had a ratchet effect
. as well, with fleeing subjects driving other peoples ahead of them seeking
safety and new territory. Much of the Southeast Asian massif is, in effect, a
shatter zone. The reputation of the southwestern Chinese province of Yun-
8 HILLS, VALLEYS, AND STATES
nan as a "museum of human races" reflects this history of migration. Shatter
zones are found wherever the expansion ofstates, empires, slave-trading, and
wars, as well as natural disasters, have driven large numbers of people to seek
refuge in out-of-the-way places: in Amazonia, in highland Latin America
(with the notable exception of the Andes, with their arable highland plateaus
and states), in that corridor of highland Africa safe from slave-raiding, in the
Balkans and the Caucasus. The diagnostic characteristics of shatter zones
are their relative geographical inaccessibility and the enormous diversity of
tongues and cultures.
Note that this account of the periphery is sharply at odds with the offi-
cial story most civilizations tell about themselves. According to that tale, a
backward, naIve, and perhaps barbaric people are gradually incorporated into
an advanced, superior, and more prosperous society and culture. If, instead,
many of these ungoverned barbarians had, at one time or another, elected,
as a political choice, to take their distance from the state, a new element'of
political agency enters the picture. Many, perhaps most, inhabitants of the
ungoverned margins are not remnants of an earlier social formation, left be-
hind, or, as some lowland folk accounts in Southeast Asia have it, "our living
ancestors." The situation of populations that have deliberately placed them-
selves at the state's periphery has occasionally been termed, infelicitously,
secondary primitivism. Their subsistence routines, their social organization,
their physical dispersal, and many elements of their culture, far from being
the archaic traits of a people left behind, are purposefully crafted both to
thwart incorporation into nearby states and to minimize the likelihood that
statelike concentrations of power will arise among them. State evasion and
state prevention permeate their practices and, often, their ideology as well.
They are, in other words, a "state effect." They are "barba:rians by design."
They continue to conduct a brisk and mutually advantageous trade with low-
land centers while steering clear of being politically captured.
Once we entertain the possibility that the "barbarians" are not just
"there" as a residue but may well have chosen their location, their subsistence
practices, and their social structure to maintain their autonomy, the standard
civilizational story of social evolution collapses utterly. The temporal, civili-
zational series-from foraging to swiddening (or to pastoralism), to sedentary
grain cultivation, to irrigated wet-rice farming-and its near-twin, the series
from roving forest bands to small clearings, to hamlets, to villages, to towns,
to court centers; these are the underpinning of the valley state's sense of su-
periority. What if the presumptive "stages" of these series were, in fact, an ar-
ray of social options, each of which represented a distinctive positioning vis-
HILLS, VALLEYS, AND STATES 9
I',
a-vis the state? And what if, over considerable periods of time, many groups
have moveq'strategically among these options toward more presumptively
"primitive" forms in order to keep the state at arm's length? On this view, the
civilizational discourse of the valley states-and not a few earlier theorists of
social evoluti9n - is not much more than a self-inflating way of confounding
the status of state-subject with civilization and that of; self-governing peoples
with primitivism.
The logic of the argument made throughout this book would essentially
reverse this logic. Most, if not all, the characteristics that appear to stigma-
tize hill peoples-their location at the margins, their physical mobility, their
swidden agriculture,their flexible social structure, their religious heterodoxy,
their egalitarianism, and even the nonliterate, oral cultures-far from being
the mark of primitives left behind by civilization, are better seen on a long
view as adaptations designed to evade both state capture and state formation.
They are,in other words, political adaptations of nonstate peoples to a world
of states that are, at once, attractive and threatening.
Creating Subjects
Avoiding the state was, until the past few centuries, a real option. A thousand
years ago most people lived outside state structures, under loose-knit empires
or in situations of fragmented sovereignty.ll Today it is an option that is fast
vanishing. To appreciate how the room for maneuver has been drastically
curtailed in the past millennium, a radically schematic and simplified fast-
forward history of the balance of power between stateless peoples and states
may be helpful.'
The permanent association of the state and sedentary agriculture is at
the center of this story,12 Fixed-field grain agriculture has been promoted
by the state and has been, historically, the foundation of its power. In turn,
sedentary agriculture leads to property rights in land, the patriarchal family
enterprise, and an emphasis, also encouraged by the state, on large families.
Grain farming is, in this respect, inherently expansionary, generating, when
not checked by disease or famine, a surplus population, which is obliged to
move and colonize new lands. By any long-run perspective, then, it is grain
agriculture that is "nomadic" and aggressive, constantly reproducing copies
of itself, while, as Hugh Brody aptly notes, foragers and hunters, relying on
, a single area and demographically far more stable, seem'by comparison "pro-
foundly settled."13
The massive expansion of European power, via colonialism and white-

10 HILLS, VALLEYS, AND STATES
settler colonies, represented a vast expansion of sedentary agriculture. In the
"neo-Europes" such as North America, Australia, Argentina, and New Z~a­
land, Europeans reproduced, as far as possible, the agriculture with which
they were familiar. In colonies with preexisting states based on sedentary
agriculture, the Europeans replaced the indigenous overlords as sovereigns,
collecting taxes and encouraging agriculture as had their predecessors, but
more effectively. All other subsistence patterns, except when they provided
valuable trade goods (for example, furs), were, fiscally speaking, considered
sterile. Thus foragers, hunters, shifting-cultivators, and pastoralists were by-
passed and ignored or driven from potentially arable farmland into territo~
ries considered wastelands. Nevertheless, as late as the end of the eighteenth
century, though they were no longer a majority of the world's population,
nonstate peoples still occupied the greater part of the world's land mass-
forest lands, rugged mountains, steppes, deserts, polar regions,. marshes, and
inaccessibly remote zones. Such regions were still a potential refuge for those
who had reason to flee the state.
These stateless peoples were not, by and large, easily drawn into the
fiscally legible economy of wage labor and sedentary agriculture. On this defi-
nition, "civilization" held little attraction for them when they could hav~ all
the advantages of trade without the drudgery, subordination, and immobility
ofstate subjects. The widespread resistance of stateless peoples led directly to
what might be called the golden age ofslavery along the littoral ofthe Atlantic
and Indian Oceans and in Southeast Asia. 14 From the perspective adopted
here, populations were forcibly removed en masse from settings where their
production and labor were illegible and inappropriable and were relocated in
colonies and plantations ,where they could be made to grow cash crops (tea,
cotton, sugar, indigo, coffee) which might contribute to the profits of land-
owners and the fiscal power of the state.ls This first step ofenclosure required
forms of capture and bondage designed to relocate them from nonstate spaces
where they were generally more autonomous (and healthy!) to places where
their labor could be appropriated.
The final two stages of this massive enclosure movement belong, in the
case of Europe, to the nineteenth century and, in the case of Southeast Asia,
largely to the late twentieth century. They mark such a radical shift in the
relationship between states and their peripheries that they fall largely outside
the story I tell here. In this last period, "enclosure" has meant not so much
shifting people from stateless zones to areas of state control but rather colo-
nizing the periphery itself and transforming it into a fully governed, fiscally
fertile zone. Its immanent logic, unlikely ever to be fully realized, is the com-
HILLS, VALLEYS, AND STATES 11
plete elimination of nonstate spaces. This·truly imperial project, made pos-
sible only by distance-demolishing technologies (all-weather roads, bridges,
railroads, airplanes, modern weapons, telegraph, telephone, and now modern
information technologies including global positioning systems), is so novel
and its dynamics so different that my analysis here makes no further sense in
,.
, . Southeast Asia for the period after, say, 1950. Modern conceptions ofnational
sovereignty and the resource needs of mature capitalism have brought that
final enclosure into view.
The hegemony, in this past century, of the nation-state as the standard
and nearly exclusive unit of sovereignty has proven profoundly inimical to
nonstate peoples. State power, in this conception, is the state's monopoly of
coercive force that must, in principle, be fully projected to the very edge of its
terdtory, where it meets, again in principle, another sovereign power project-
ing its command to its own adjacent frontier. Gone, in principle, are the large
areas of no sovereignty or mutually canceling weak sovereignties. Gone too,
of course, are peoples under no particular sovereignty. As a practical matter,
most nation-states have tried, insofar as they had the means, to give substance
to this vision, establishing armed border posts, moving loyal populations to
the frontier and, relocating or driving away "disloyal" populations, clearing
frontier lands for sedentary agriculture, building roads to the borders, and
registering hitherto fugitive peoples.
On the heels of this notion of sovereignty came the realization that
these neglected and seemingly useless territories to which stateless peoples
had been relegated were suddenly of great value to the economies of mature
capitalism.16 They contained valuable resources-oil, iron ore, copper, lead,
timber, uranium, bauxite, the rare metals essential to the aerospace and
electronics industries, hydroelectric sites, bioprospecting and conservation
! '.
J, areas-that might in many cases be the linchpin of state revenue. Places that
long ago might have been desirable for their deposits ofsilver, gold, and gems,
not to mention slaves, became the object of a new gold rush. All the more
!, reason to project state power to the nethermost reaches of these ungoverned
regions and bring their inhabitants under firm control.
Occupying and controlling the margins of the state implied a cultural
policy as well. Much of the periphery along national borders of mainland
Southeast Asia is inhabited by peoples linguistically and culturally distinct
from the populations that dominate the state cores. Alarmingly, they spill
.promiscuously across national frontiers, generating multiple identities and
possible foci of irredentism or secession. Weak valley states have permitted,
or rather tolerated, a certain degree ofautonomy when they had little choice.
 ."'
12 HILLS, VALLEYS, AND STATES
Where they could, however, all states in the region have tried to bring s'Uch
peoples under their routine administration, to encourage and, more rarely,
to insist upon linguistic, cultural, and religious alignment with the majority
population at the state core. This meant, in Thailand, encouraging, say, the
Lahu to become Thai-speaking, literate, Buddhist subjects of the monarchy.
In Burma it meant encouraging, say, the Karen to become Burmese-speaking
Buddhists loyal to the military juntaP
Parallel to policies of economic, administrative, and cultural absorption
has been the policy, driven by both demographic pressure and self-conscious
design, of ·engulfment. Huge numbers of land-hungry majorities from the
plains have moved, or been moved, to the hills. There, they replicate valley
settlement patterns and sedentary agriculture, and, over time, they de~o­
graphically dominate the dispersed, less numerous hill peoples. The com-
bination of forced settlement and engulfment is nicely illustrated by a series
of Vietnamese mobilization campaigns in the 1950S and 1960s: "Campaign
to Sedentarize the Nomads," "Campaign for Fixed Cultivation and Fixed
Residence," "Storm the Hills Campaign," and "Clear the Hills by Torchlight
Campaign. "18
Culturally, this reduction and standardization of relatively autonomous,
self-governing communities is a process of long historical lineage. It is an
integral theme of the historical consciousness of each of the large mainland
Southeast Asian states. In the Vietnamese official national narrative, the
"march to the south" -to the Mekong and the trans-Bassac Deltas-inaccu-
rate though it is as a description of the historical process, vies with the wars of
national liberation for pride of place. 19 Burmese and Thai history are no less
marked by the movement of population from their more northern historical
cores of Mandalay, Ayutthaya, and what is now Hanoi into the Irrawaddy,
Chao Praya, and Mekong river deltas, respectively. The great cosmopolitan,
maritime cities of Saigon (now Ho Chi Minh City), Rangoon, and Bangkok
that grew to serve this onetime frontier, delta, hinterland have come, demo-
graphically, to dominate the earlier inland capitals.
Internal colonialism, broadly understood, aptly describes this process.
It involved the absorption, displacement, and/or extermination of the previ-
ous inhabitants. It involved a botanical colonization in which the landscape
was transformed - by deforestation, drainage, irrigation, and levees - to ac-
commodate crops, settlement patterns, and systems of administration famil-
iar to the state and to the colonists. One way of appreciating the effect of this
colonization is to view it as a massive reduction of vernaculars of all kinds: of
vernacular languages, minority peoples, vernacular cultivation techniques,
HILLS, VALLEYS, AND STATES 13
vernacular land tenure systems, vernacular hunting, gathering, and forestry
techniques, vernacular religion, and so on. The attempt to bring the periph-
ery into line is read by representatives of the sponsoring state as provid-
ing civilization and progress-where progress is, in turn, read as the intru-
sive propagation of the linguistic, agricultural, and religious practices of the
dominant ethnic group: the Han, the Kinh, the Burm<itn, the Thai. 2o
The remaining self-governing peoples and spaces of mainland South-
east Asia are much diminished. We shall, for the most part, concentrate on
the so-called hill peoples (often mistakenly called tribes) of mainland South-
east Asia, particularly Burma. While I will clarify what I mean by the awk-
ward term nonstate spaces, it is not simply a synonym for hills or for higher
altitudes. States, being associated with concentrated grain production, typi-
cally arise where there is a substantial expanse of arable land. In mainland
Southeast Asia, this agro-ecology is generally at low elevations, allowing us
to speak of "valley states" and "hill peoples." Where, as in the Andes, most
easily cultivable land under traditional conditions is located at high eleva-
tions, it is the other way around. The states were in the hills and nonstate
spaces were downhill in the humid lowlands. Thus the key variable is not
so much elevation per se as the possibility for concentrated grain produc-
tion. Nonstate space, by contrast, points to locations where, owing largely to
geographical obstacles, the state has particular difficulty in establishing and
maintaining its authority. A Ming emperor had something like this in mind
when he described the southwest provinces of his kingdom: "The roads are
long and dangerous, the mountains and rivers present great obstacles, and
the customs and practices differ. "21 But swamps, marshes, mangrove coasts,
deserts, volcanic margins, and even the open sea, like the' ever growing and
changing deltas of Southeast Asia's great rivers, all function in much the
same way. Thus it is difficult or inaccessible terrain, regardless of elevation,
that presents great obstacles to state control. As we shall see at great length,
such places have often served as havens of refuge for peoples resisting or flee-
ing the state.
The Great Mountain Kingdom; or, "Zomia"; or,
The Marches of Mainland Southeast Asia
One of the largest remaining nonstate spaces in the world, if not the largest,
·is the vast expanse of uplands, variously termed the Southeast Asian mas-
sif and, more recently, Zomia.22 This great mountain realm on the marches
14 HILLS, VALLEYS, AND STATES

of mainland Southeast Asia, China, India, and Bangladesh sprawls across

roughly 2.5 million square kilometers-an area roughly the size of Europe.
As one of the first scholars to identify the massif and its peoples as a single
object of study, Jean Michaud has traced its extent: "From north to south, it
includes southern and western Sichuan, all of Guizhou and Yunnan, western
and northern Guangxi, western Guangdong, most of northern Burma with
an adjacent segment of extreme [north]eastern India, the north and west of
Thailand, practically all of Laos above the Mekong Valley, northern and cen-
tral Vietnam along the Annam Cordillera, and the north and eastern fringes
of Cambodia. "23
Rough calculations would put Zomia minority populations alone at
around eighty million to one hundred million. 24 Its peoples are fragment~d
into hundreds of ethnic identities and at least five language families that defy
any simple classification.
Lying at altitudes from two hundred or three hundred'meters above
sea level to more than four thousand meters, Zomia could be thought of as
a Southeast Asian Appalachia, were it not for the fact that it sprawls across
eight nation-states. A better analogy would be Switzerland, a mountain
kingdom at the periphery of Germany, France, and Italy that itself became I
'If.
a nation-state. Borrowing Ernest Gellner's felicitous phrase referring to the
Berbers of the High Atlas Mountains, this huge hilly zone might be seen as a
"pervasive Switzerland without cuckoo clocks. "25 Far from being a hilly na-
tion, however, this upland belt lies on the marches, far from the main popu- I
i,
lation centers of the nations it traverses. 26 Zomia is marginal in almost every
respect. It lies at a great distance from the main centers of economic activity;
it bestrides a contact zone between eight nation-states and several religious
traditions and cosmologies. 27
I
Scholarship organized historically around the classical states and their I;

cultural cores and, more recently, around the nation-state is singularly ill-
equipped to examine this upland belt as a whole. Willem van Schendel is
one of a handful of pioneers who. have argued that these cumulative nation-
state "shards" merit consideration as a distinctive region. He has gone so far
as to give it the dignity of a name of its own: Zomia, a term for highlander
common to several related Tibeto-Burman languages spoken in the India-
Bangladesh-Burma border area. 28 More precisely, Zo is a relational term
meaning "remote" and hence carries the connotation of living in the hills;

Map I. Mainland Southeast Asia

16 HILLS, VALLEYS, AND STATES

Mi means "people." As is the case elsewhere in Southeast Asia Mi-zo or

Zo-mi designated a remote hill people, while at the same time the ethnic label
applies to a geographical niche. 29 Although van Schendel proposes a bold.
expansion of Zomia's boundaries to Afghanistan and beyond, I will confine
my use of the term to the hilly areas eastward, beginning with the Naga and
Mizo hills in northern India and Bangladesh's Chittagong Hill Tracts.
Zomia, at first glance, would seem an unlikely candidate for consider-
ation as a distinctive region. The premise for calling a geographical area a:
region is typically that it shares important cultural features that mark it of!
,i,
from adjacent areas. In this fashion, Fernand Braudel was able to show that
the coastal societies around the Mediterranean Sea constituted a region,
owing to their long and intense commercial and cultural connections. 30 De-
spite political and religious chasms between, say, Venice and Istanbul, they
were integral parts ofa recognizable world of exchange and mutual influence.
Anthony Reid has made a similar, and in many respects, more powerful claim
for the Sunda Shelf littoral in maritime Southeast Asja, where trade and mi-
gration were, if anything, easier than in the Mediterranean. 31 The principle
behind region-making in each case is that, for the premodern world, water,
especially if it is calm, joins people, whereas mountains, especially if they are
high and rugged, divide people. As late as 1740 it took no more time to s~il
from Southampton to the Cape of Good Hope than to travel by stagecoach
from London to Edinburgh.
On these grounds, hilly Zomia would seem to be a "negative" region.
Variety, more than uniformity, is its trademark. In the space of a hundred
kilometers in the hills one can find more cultural variation - in language,
dress, settlement pattern, ethnic identification, economic activity, and reli-
gious practices - than one would ever find in the lowland river valleys. Zomia
may not quite attain the prodigious cultural variety of deeply fissured New
Guinea, but its complex ethnic and linguistic mosaic has presented a be-
wildering puzzle for ethnographers and historians, not to mention would-be
rulers. Scholarly work on the area has been as fragmented and isolated as the
terrain itself seemed to be. 32
I will argue not only that Zomia qualifies as a region in the strong sense
of the term, but also that it is impossible to provide a satisfactory account of
the valley states without understanding the central role played by Zomia in
their formation and collapse. The dialectic or coevolution of hill and valley,

Map 2. "Zqmia," on the mainland Southeast Asian massif

18 HILLS, VALLEYS, AND STATES
as antagonistic but deeply connected spaces, is, I believe, the essential point
of departure for making sense of historical change in Southeast Asia. . distances. The "Akha" along the Yunnan-Thai border and the "Hani" in the
Most of what the hills share as physical and social spaces marks them off
fairly sharply from the more populous lowland centers. The population ofthe
hills is far more dispersed and culturally diverse than that of the valleys. It is
as if the difficulties of terrain and relative isolation have, over many centuries,
encouraged a kind of "speciation" of languages, dialects, dress, and cultural
practices. The relative availability of forest resources and open, if steep, land
has also allowed far more diverse subsistence practices than in the valleys,
where wet-rice monocropping often prevails. Swiddening (or slash-and-burn
agriculture), which requires more land and requires clearing new fields and
occasionally shifting settlement sites, is far more common in the hills.
As a general rule, social structure in the hills is both more flexible and
more egalitarian than in the hierarchical, codified valley societies. Hybrid
identities, movement, and the social fluidity that characterizes many frontier
societies are common. Early colonial officials, taking an inventory of their
new possessions in the hills, were confused to encounter hamlets with several
"peoples" living side by side: hill people who spoke three or four languages
and both individuals and groups whose ethnic identity had shifted, some-
times within a single generation. Aspiring to Linnaean specificity in the clas-
sification of peoples as well as flora, territorial administrators were constantly
frustrated by the bewildering flux of peoples who refused to stay put. There
was, however, one principle of location that brought some order to thisap-
parent anarchy of identity, and that was its relation to altitude. 33 As Edmund
Leach originally suggested, once one looks at Zomia not from a high-altitude
balloon but, rather, horizontally, in terms of lateral slices through the topog-
raphy, a certain order emerges.34 In any given landscape, particular groups
often settled within a narrow range of altitudes to exploit the agro-economic
possibilities of that particular niche. Thus, for example, the Hmong :have
tended to settle at very high altitudes (between one thousand and eighteen
hundred meters) and to plant maize, opium, and millet that will thrive at
that elevation.' If from a high-altitude balloon or on a map they appear to
be a random scattering of small blotches, this is because they have occupied
the mountaintops and left the midslopes and intervening valleys to other
groups.
Specialization by altitude and niche within the hills leads to scattering.
And yet long-distance travel, marriage alliances, similar subsistence patterns,
and cultural continuity help foster coherent identities across considerable
HILLS, VALLEYS, AND STATES 19
upper reaches of the Red River in northern Vietnam are recognizably the
same culture, though separated by more than a thousand kilometers. They
, typically have more in common with each another than either group has with
valley people a mere thirty or forty miles away. Zomia is thus knitted together
as a region not by a political unity, which it utterly lacks, but by compa-
rable patterns of diverse hill agriculture, dispersal and mobility, and rough
egalitarianism, which, not incidentally, includes a relatively higher status for
women than in the valleys.35
The signal, distinguishing trait of Zomia, vis-a.-vis the lowland regions
it borders, is that it is relatively stateless. Historically, of course, there have
been states in the hills where a substantial fertile plateau and/or a key node
in the overland trade routes made it possible. Nan Chao, Kengtung, Nan, and
Lan-na were among the best known. 36 They are the exceptions that prove the
rule. While state-making projects have abounded in the hills, it is fair to say
I,
that few have come to fruition. Those would-be kingdoms that did manage
to defy the odds did so only for a relatively brief, crisis-strewn period.
i· Such episodes aside, the hills, unlike the valleys, have paid neither taxes
to monarchs nor regular tithes to a permanent religious establishment. They
have constituted a relatively free, stateless population of foragers and hill
farmers. Zomia's situation at the frontiers of lowland state centers has con-
tributed to its relative isolation and the autonomy that such isolation favors.
Lying athwart state borders where multiple competing sovereignties abut
one another has itself afforded its peoples certain advantages for smuggling,
contraband, opium production, and the "small border powers" that negotiate
a tenuous, high-wire act of quasi-independence. 37
j,
'
A stronger and, I believe, more accurate political description is that the
hill populations of Zomia have actively resisted incorporation into the frame-
work ofthe classical state, the colonial state, and the independent nation-state.
Beyond merely taking advantage of their geographical isolation from centers
of state power, much of Zomia has "resisted the projects of nation'-building
and state-making of the states to which it belonged."38 This resistance came
especially to light after the creation of independent states alter World War II,
when Zomia became the site of secessionist movements, indigenous rights
struggles, millennial rebellions, regionalist agitation, and armed opposition
to lowland states. But it is a resistance with deeper roots. In the precolonial
. period, the resistance can be seen in a cultural refusal oflowland patterns and
in the flight of lowlanders seeking refuge in the hills.
20 HILLS, VALLEYS, AND STATES
During the colonial era, the autonomy of the hills, politically and cul-
turally, was underwritten by the Europeans for whom a separately adminis,:"
tered hill zone was a makeweight against the lowland majorities resentful of
colonial rule. One effect of this classic divide-and-rule policy is that, with a
few exceptions, hill peoples typically played little or no role-or an ,antago-
nistic one - in the anticolonial movements. They remained, at best, marginal
to the nationalist narrative or, at worst, were seen as a fifth column threat.:..
ening that independence. It is partly for such reasons that the postcolonial
lowland states have sought fully to exercise authority in the hills: by military
occupation, by campaigns against shifting cultivation, by forced settlements,
by promoting the migration of lowlanders to the hills, by efforts at religious
conversion, by space-conquering roads, bridges, and telephone lines, and by
development schemes that project government administration, and lowland
cultural styles into the hills.
The hills, however, are not simply a space of political resistance but
also a zone of cultural refusal. If it were merely a matter of political authority,
one might expect the hill society to resemble valley society culturally except
for their altitude and the dispersed settlement that the terrain favors. But
the hill populations do not generally resemble the valley centers culturally,
religiously, or linguistically. This cultural chasm between the mountains and
the plains has been claimed as something of a historical constant in Europe
as well, until quite recently. Fernand Braudel acknowledged the political au-
tonomy of the hills when he approvingly quoted Baron de Tott to the effect
that "the steepest places have always been the asylum of liberty." But he
carried the argument much further, asserting the existence of an unbridge-
able cultural gap between plains and mountains. He wrote: "The mountains
are as a rule a world apart from civilizations which are an urban and lowland
achievement. Their history is to have none, to remain always on the fringes
of the great waves of civilization, even the longest and most persistent, which
may spread over great distances in the horizontal plane but are powerless to
move vertically when faced with an obstacle of several hundred meters."39
Braudel was, in turn, only echoing a much older view captured by the great
fourteenth...,century Arab philosopher Ibn Khaldun, who noted that "Arabs
can gain control only over flat territory" and do not pursue tribes that hide
in the mountains.40 Compare Braudel's bold assertion that civilizations can't
climb hills to a nearly identical assertion made by Oliver Wolters, quoting
Paul Wheatley, about precolonial Southeast Asia: "Many people lived in the
distant highlands and were beyond the reach of the centers where records
HILLS, VALLEYS, AND STATES 21
survive. The mandalas [court centers of civilization and power] were a phe-
nomenon of the lowlands and even there, geographical conditions encour-
aged under-government. Paul Wheatley puts it well when he notes that 'the
Sanskritic tongue was stilled to silence at 500 meters.' "41
Scholars of Southeast Asia have been struck again and again by the
sharp limits the terrain, particularly altitude, has placed on cultural or po-
litical influence. Paul Mus, writing of Vietnam and ~choing Wheatley, noted
of the spread of the Vietnamese and their culture that "this ethnic adven-
ture stopped at the foot of the high country's buttresses."42 Owen Latti-
more, best known for his studies of China's northern frontier, also remarked
that Indian and Chinese civilizations, like those cited by Braudel, traveled
well across the plains but ran out of breath when they encountered rugged
hills: "This kind of stratification extends far beyond China itself into the
Indochinese peninsula, Thailand and Burma with the influence of the an-
cient high civilizations reaching far out over the lower levels where concen-
i,
trated agriculture and big cities are to be found, but not up into the higher
altitudes."43
Though Zomia is exceptionally'diverse linguistically, the languages spo-
ken in the hills are, as a rule, distinct from those spoken in the plains. Kinship
structures, at least formally, also distinguish the hills from the lowlands. This
is in part what Edmund Leach had in mind when he characterized hill society
'as following a "Chinese model" while lowland society followed an "Indian"
or Sanskritic model.44
'Hill societies are, as a rule, systematically different from valley soci-
eties. Hill people tend to be animists, or, in the twentieth century, Christians,
~,
\
who do not follow the "great tradition" salvation religions oflowland peoples
(Buddhism and Islam in particular). Where, as occasionally happens, they do
come to embrace the "world religion" of their valley neighbors, they are likely
to do so with a degree of heterodoxy and millenarian fervor that valley elites
:'1 find more threatening than reassuring. Hill societies do produce a surplus,
but they do not use that surplus to support kings and monks. The absence
of large, permanent, surplus-absorbing religious and political establishments
makes for a sociological pyramid in the hills that is rather flat and local when
compared with that of valley societies. Distinctions of status and wealth
abound in the hills, as in the valleys. The difference is that in the valleys they
tend to be supralocal and enduring, while in the hills they are both unstable
and geographically confined.
This characterization obscures a great deal of variation in the political
22 HILLS, VALLEYS, AND STATES
structure of hill societies. The variation is not by any means simply a func-
tion of "ethnicity," although some hill peoples, such as the Lahu, Khmu,
and Akha, seem strongly egalitarian and decentralized. It is just as common,
however, to encounter groups that defy such 'generalizations. Among Karen,
Kachin, Chin, Hmong, Yao/Mien, and Wa, for example, there seem to be
both relatively hierarchical subgroups and relatively decentralized, egalitar-
ian subgroups. What is most striking and important is that the degree of
hierarchy and centralization is not constant over time. The variation, so far
as I can make out, depends largely on a kind of imitative state-making. That
is, it is either a kind of short-term war alliance or a sort of "booty-capitalism"
for slave-raiding and extracting tribute from lowland communities. Where
hill groups are in a tributary relationship with a valley kingdom-which ,does
not imply political incorporation or, necessarily, inferiority-it may be -an
expedient to control a lucrative trade route or to safeguard privileged access
to valuable markets. Their political structures are, with extremely rare excep-
tions, imitative in the sense that while they may have the trappings and rhe-
toric of monarchy, they lack the substance: a taxpaying subject population or
direct control over their constituent units, let alone a standing army. Hill poli-
ties are, almost invariably, redistributive, competitive feasting systems held
together by the benefits they are able to disburse. When they occasionally
appear to be relatively centralized, they resemble what Barfield has called the
"shadow-empires" of nomadic pastoralists, a predatory periphery desigp.ed
to monopolize trading and raiding advantages at the edge of an empire. They
are also typically parasitic in the sense that when their host-empires collapse,
so do they.45
; ,
Zones of Refuge
There is strong evidence that Zomia is not simply a region of resistan~e to
valley states, but a region of refuge as well. 46 By "refuge," I mean to imply
that much of the population in the hills has, for more than a millennium ,and
a half, come there to evade the manifold afflictions of state-making projects
in the valleys. Far from being "left behind" by the progress of civilization in
the valleys, they have, over long periods of time, chosen to place themselves
out of the reach of the state. Jean Michaud notes, in this connection, that
what he calls nomadism in the hills can be "an escape or survival strategy"
and sees the unprecedented series of massive rebellions in the latter half of
the nineteenth century in central and southwest China as having pushed the
HILLS, VALLEYS, AND STATES 23
millions of refugees streaming south into the more remote highlands. He is
sympathetic to the view adopted here that Zomia is best seen historically as
a region of refuge from states, most especially the Han state. "It is probably
fair to say," he concludes, "that the highland populations who migrated from
China to the ... highlands over the past five centuries were, at least in part,
pushed from their homelands by aggression from more powerful neighbors,
including especially Han expansion."47
Detailed and unambiguous documentary evidence of the conflicts gen-
erated by Han expansion and the migratory flights it provoked is abundant
from the early Ming Dynasty (1368) onward, becoming even more abundant
under the Q!ng. Earlier documentation is harder to come by and more am-
biguous, owing to the great fluidity of ethnic and political labels. The general
pattern, l!owever, seems to be as follows: as the reach of the Chinese state
grew, peoples at the point of expansion were either absorbed (becoming, in
time, Han) or moved away, often after a failed revolt. Those who left be-
came, at least for a time, distinct societies that could be said to have "self-
marginalized" by migration.48 As the process was repeated again and again,
culturally complex zones of refuge sprang up in the hinterlands of the state.
"The history of the various non-state peoples of this region" can, Fiskesjo
believes, be written as the bifurcation between those who had long been in
the hills (for example, the Wa people) and those who sought refuge there:
"Among those who left [the zone of Chinese state power], we find many
Tibeto-Burman ethnolinguistic formations (Lahu, Hani, Akha, etc.) as well
as Miao or Hmong speakers, and other peoples . . . described as 'hill tribes
out of China' with a 'heritage of defeat' that has led many of them during
the past few centuries, into the northern parts of the modern states of Thai-
land, Burma, Laos, and Vietnam where many of them are still regarded as
newcomers."49
There, in regions beyond the states' immediate writ and, thus, at some
remove from taxes, corvee labor, conscription, and the more than ,occasional
epidemics and crop failures associated with population concentration' and
monocropping, such groups found relative freedom and safety. There, they
practiced what I will call escape agriculture: forms of cultivation designed to
thwart state appropriation. Even their social structure could fairly be called
escape social structure inasmuch as it was designed to aid dispersal and au-
tonomy and to ward off political subordination.
The tremendous linguistic and ethnic fluidity in the hills is itself a cru-
cial social resourc~ for adapting to changing constellations of power, inas-
 ; ,
24 HILLS, VALLEYS, AND STATES
much as it facilitates remarkable feats of identity shape-shifting. Zomians
are not as a rule only linguistically and ethnically amphibious; they are, in
their strong inclination to follow charismatic figures who arise among them,
capable of nearly instantaneous social change, abandoning their fields and
houses to join or form a new community at the behest of a trusted prophet.
Their capacity to "turn on a dime" represents the ultimate in escape sociaJ
structure. Illiteracy in the hills can, more speculatively, be interpreted in the
same fashion. Virtually all hill peoples have legends claiming that they once
had writing and either lost it or that it was stolen from them. Given the con- ,
siderable advantages in plasticity of oral over written histories and genealo-
gies, it is at least conceivable to see the loss of literacy and of written texts as
a more or less deliberate adaptation to statelessness. . less than runaway serfs from all over European Russia who accumulated at
The argument, in short, is that the history of hill peoples is best under-
stood as a history not of archaic remnants but of "runaways" from state-
making processes in the lowlands: a largely "maroon" society, providing'
that we take a very long historical view. Many of the agricultural and so-
cial practices of hill peoples can be best understood as techniques to make
good this evasion, while maintaining the economic advantages of the lowland "
connection.
The concentration of people and production at a single location -re-
quired some form of unfree labor when population was sparse, as it was in
Southeast Asia. All Southeast Asian states were slaving states, without eicep-
tion, some of them until well into the twentieth century. Wars in precolonial
Southeast Asia were less about territory than about the seizure of as many
captives as possible who were then resettled at the core of the winner's terri-
tory. They were not distinctive in this respect. After all, in Periclean Athens,
the population of slaves outnumbered full citizens by five to one.
The effect of aJI state-making projects of this kind was to create a shat-
ter zone or flight zone to which those wishing to evade or to escape bondage
fled. These regions of refuge constituted a direct "state effect." Zomia simply
happens to be, owing largely to the precocious early expansion of the Chinese
state, one of the most extensive and oldest zones of refuge. Such regions are,
however, inevitable by-products of coercive state-making and are found on
every continent. A few of them will figure as comparative cases in what fol-
lows, but here I want to enumerate several examples to suggest how common
they are.
The forced-labor characteristic of Spanish colonization in the New
World provoked the widespread flight of native peoples out of range, often
HILLS, VALLEYS, AND STATES 25
to hilly or arid places where they could live unmolested. 50 Such areas were
marked by great linguistic and ethnic diversity and occasionally by a simpli-
fication of social structure and subsistence routines-foraging, shifting culti-
vation-to increase mobility. The process was repeated in the Spanish Phil-
ippines, where, it is claimed, the cordillera of northern Luzon was populated
almost entirely by lowland Filipinos fleeing Malay slave raids and the Span-
ish reducciones. 51 As peoples adapted to hill ecology, a process of ethnogenesis
followed, after which highland Filipinos were later misrepresented as the
descendants of separate, prehistoric migrations to the island.
The Cossacks on Russia's many frontiers represent another striking
example of the process. They were, at the outset, nothing more and nothing
the fronti 7r.52 They became, depending on their location, different Cossack
"hosts": the Don (for the Don River basin) Cossacks, the Azov (Sea) Cos-
sacks, and so on. There at the frontier, copying the horseback habits of their
Tatar neighbors and sharing a common open...land pasture, they became "a
people," later used by the tsars, the Ottomans, and'the Poles as cavalry. The
history of the Roma and Sinti (Gypsies) in late-seventeenth-century Europe
provides a further striking example. 53 Along with other stigmatized itinerant
peoples, they were' subject to two forms of penal labor: galley slavery in the
Mediterranean basin and, in the northeast,""forced.conscription as soldiers or
military porters in Prussia-Brandenburg. As a result they accumulated in a
narrow band of territory that came to be known as the "outlaw corridor," the
one location between the catchment areas ofthese twin, mortal dangers.
Inasmuch as the captivity and bondage associated with early state-
making generate, in their wake, flight and zones of refuge, slavery as a labor
system produced many "Zomias',' large and small. It is possible, in this con-
text, to delineate an upland, remote zone of West Africa that was relatively
safe from the five hundred-year-long worldwide slave-raiding and trade that
caught tens of millions of in its toils. 54 This zone of refuge grew it!. popula-
tion despite the difficulties of the terrain and the necessity for new subsis-
tence routines. Many of those who failed to evade the slave raids in Africa,
once transplanted to the New World, promptly escaped and created fugitive
slave (maroon) settlements wherever slavery was practiced: the famous high-
land "cockpit" ofJamaica; Palmares in Brazil, a maroon community of some
twenty thousand inhabitants; and Surinam, the largest maroon population in
the hemisphere, are only three illustrations. Were we to include smaller scale
"refugia" such as marshes, swamps, and deltas, the list would multiply many-

26 HILLS, VALLEYS, AND STATES
fold. To mention only a few, the great marsh on the lower Euphrates (drained
under Saddam Hussein's rule) was for two thousand years a refuge from state
control. So, on a smaller scale, were the storied Great Dismal Swamp on the
North Carolina-Virginia border, the Pripet Marshes in Poland, now on the
Belarus-Ukraine border, and the Pontian Marshes near Rome (drained finally
by Mussolini) known as zones of refuge from the state. The ~ist ~f such refu~
gia is at least as long as the list of coercive labor schemes that mevltably spawn
them.
Hill societies in mainland Southeast Asia, then, for all their riotous
heterogeneity, have certain characteristics in common, and most of these
characteristics distinguish them sharply from their valley neighbors. Th~y
encode a pattern of historic flight and hence a position of opposition if riot
resistance. If it is this historical, structural relation that we hope to illuminate,
then it makes no sense whatever to confine ourselves to a nation-state frame-
work. For much of the period we wish to examine there was no nation-state
and, when it did come into being late in the game, many hill people continued
to conduct their cross-border lives as if the state didn't exist. The concept
of "Zomia" marks an attempt to explore a new genre 0 f " area" stu d'les, m
which the justification for designating the area has nothing to do with national
boundaries (for example, Laos) or strategic conceptions (for example, South-
east Asia) but is rather based on certain ecological regularities and structu~al
relationships that do not hesitate to cross national frontiers. If we have our
way, the example of "Zomia studies" will inspire others to follow this experi-
ment elsewhere and improve on it.
The Symbiotic History of Hills and Valleys
Histories of the classical lowland court-states, taken in isolation, risk being
unintelligible or vastly misleading. Lowland states (mandala or modern) have
always existed in symbiosis with hill society. 55 By symbiosis, I mean to invoke
the biological metaphor of two organisms living together in more or less inti-
mate association - in this case, social organisms.' The term does not specify,
nor do I wish to do so here, whether this mutual dependence is antagonistic,
or even parasitic, or whether it is mutually beneficial, "synergistic."
It is not possible to write a coherent history of the hills that is not in
constant dialogue with lowland centers; nor is it possible to write a coher-
ent history of lowland centers that ignores its hilly periphery. By and large,
most students of hill societies have been sensitive to this dialectic, stressing
HILLS, VALLEYS, AND STATES 27
the deep history of symbolic, economic, and human traffic between the two
societies. The same typically cannot be said of work - even the most dis-
tinguished-on lowland centers. 56 The pattern is hardly surprising. Treat-
ment oflowland cultures and societies as self-contained entities (for example,
"Thai civilization," "Chinese culture") replicates the unreflective structure
of scholarship and, in doing so, adopts the hermetic v~ew of culture that low-
land elites themselves wish to project. The fact is that'hill and valley societies
have to be read against each other to make any sense. I attempt just such a
reading here.
Writing an account of valley population centers without including the
hills would be like writing a history of colonial New England and the Middle
Atlantic States without considering the American frontier. It would be like
writing a history of antebellum slavery in the United States while leaving out
the freedmen and the lure of freedom in Canada. In each case, an external
frontier conditioned, bounded, and in many respects constituted what was
possible at the center. Accounts of lowland states that miss this dimension do
not merely "leave out" the hills; they ignore a set of boundary conditions and
exchanges that make the center what it is.
. The constant movement back and forth between the valleys and the
hills-its causes, its patterns, its consequences-will preoccupy us. Many
:~ ,;
valley people are, as it were, "ex-hill people," and many hill people are "ex-
valley people." Nor did movement in one direction or the other preclude
subsequent moves. Depending on the circumstances, groups have disengaged
themselves from a state and then, later, sought to affiliate themselves (or been
seized by!) the same or another state. A century or two later, they might again
be found outside that state's grasp, perhaps because they had moved away or
perhaps because the state in question had itself collapsed. Such shifts were
often accompanied by a shift in ethnic identity, broadly understood. I will ar-
gue for a radically "constructionist" understanding of the so-called hill tribes
of mainland Southeast Asia. They are best understood, at least as a first ap-
proximation, as a fugitive population that has come to the hills ovef'the past
two millennia. This flight was not only from the Burman, Tai, and Siamese
states but also, and most especially, from the Han Empire during the expan-
sionary phases of the Tang, Yuan, Ming, and Q!ng dynasties, when its forces
and settlers pressed into southwest China. In the hills they might have moved
several times subsequently, pressed by other, stronger fugitives or threatened
by a new state expansion, or in search of new land and autonomy. Their loca-
tion and many of their economic and cultural practices could again fairly be

28 HILLS, VALLEYS, AND STATES
termed a state effect. This picture is radically at odds with older prevailing
assumptions of a primeval population in the hills abandoned by those who
moved downhill and developed civilizations. . rights of segmentary societies in opposition to the organs of state power."
By the same token, the valley centers of wet-rice cultivation may prof-
itably be seen as constituting a hill effect in the following ways. The val-
ley states are, of course, new structures historically speaking, dating back to
roughly the middle of the first millennium CEo They were formed from ali.
earlier ingathering of diverse peoples, some of whom may have adopted fixed-
field agriculture, but who were, by definition, not previously part ofan estab-
lished state.57 The very earliest mandala states were less engines of military
conquest than cultural spaces available to all those who wished to conform
to their religious, linguistic, and cultural formats, whatever their origip.58
Perhaps because such identities were newly confected from many cultural
shards, the resulting valley self-representations were at pains to distinguish
their culture from populations outside the state. Thus if hill society could be
termed a state effect, valley culture could be seen as a hill effect.
Most of the terms that we would translate as crude, unrefined, barbaric,
and, in the Chinese case, raw refer directly to those who live in the hills
and forests. "Forest dweller" or "hill person" is shorthand for "uncivilized."
Thus, despite a centuries-old, brisk traffic in people, goods, and culture across
the very permeable membrane between the hills and valleys, it is striking how
stark and durable the cultural divide remains in lived experience. Valley and
hill peoples generally have an essentialist understanding of the differences
between them that appears to be at odds with the historical evidence over the
long run.
How can we make sense of this paradox? Perhaps the first step is to em-
phasize that the relationship between valley states and hill society is not just
symbiotic but also both contemporaneous and quasi-oppositional. In older
understandings of hill "tribes," not to mention popular folklore today, they
are considered to be the historical remnants of an earlier stage of human his-
tory: what we were like before we discovered wet-rice agriculture, learned to
write, developed the arts of civilization, and adopted Buddhism. While this
"just-so" story treats valley cultures as later, and higher, achievements of
civilization, raised from the muck of tribalism, as it were, it grossly distorts
the historical record. Valley states and hill peoples are, instead, constituted
in each other's shadow, both reciprocal and contemporaneous. Hill societies
have always been in touch with imperial states in the valleys directly or via
maritime trade routes. Valley states, by the same token, have always been in
HILLS, VALLEYS, AND STATES 29
touch with the nonstate periphery - what Deleuze and Guattari call "the
local mechanisms of bands, margins, minorities, which continue to affirm the
Such states are, in fact, "inconceivable independent of that relationship."59
Precisely the same case has been made about the relationship between
itinerant peoples-including pastoral nomads-and s~ates. Thus Pierre Clas-
tres argues persuasively that the so-called primitive Amerindian societies of
South America were not ancient societies that had failed to invent settled
agriculture or state forms but rather previously sedentary cultivators who
abandoned agriculture and fixed villages in response to the effects of the
Conquest: both disease-induced demographic collapse and colonial forced
labor.60 Their movement and subsistence techniques were designed to ward
off incorporation into the state. On the steppes of Central Asia the most an-
cient nomads, Griaznov has shown, were former sedentary cultivators who
similarly left cultivation behind for political and demographic reasons. 61 Lat-
timore reached the same conclusion, insisting that pastoral nomadism arose
after farming and drew in sedentary cultivators at the edge of the grasslands
who "had detached themselves from farming communities."62 Far from
being successive stages in social evolution, such states and nomadic peoples
are twins, born more or less at the same time and joined in a sometimes ran-
corous but unavoidable embrace.
, This pattern of paired symbiosis and opposition is a staple of Middle
Eastern history and anthropology. In the Maghreb it takes the form of struc-
tural opposition between Arabs and Berbers. Ernest Gellner's classic Saints
of the Atlas captures the dynamic I have in mind. Gellner, too, emphasized
that the political autonomy and tribalism of the Berber population in the
'High Atlas is "not a tribalism 'prior to government' but a political and par-
tial rejection of a particular government combined with some acceptance of
a wider culture and its ethic. "63 Sharing elements of a larger culture and a
faith in Islam, such tribal opposition is explicitly political and deliberately
so. Until very recently, Gellner claims, Moroccan history could be 'written in
terms of the opposition between the land of makhazen (the pale) and the land
of siba (beyond the pale). Siba could be defined as "institutional dissidence,"
though it has sometimes been translated as "anarchy." In practice, siba means
"ungoverned," a zone of political autonomy and independence, while makha-
zen means "governed," subordinated to the state. Political autonomy was,
Gellner insists, a choice, not a given.
To those groups that have self-consciously elected to move or to stay
30 HILLS, VALLEYS, AND STATES
beyond the pale, Gellner applies the term marginal tribalism to emphasize that
their marginality is a political stance:
Such tribesmen know the possibility ... of being incorporated in a more cen-
tralized state.... Indeed, they may have deliberately rejected and violently
resisted the alternative. The tribes of the High Atlas ,are of this kind. Until the
advent of the modern state, they were dissident and self-consciously so. . . .
"Marginal" tribalism ... [is] the type of tribal society which exists at the edge
of non-tribal societies. It arises from the fact that the inconveniences of sub-
mission make it attractive to withdraw from political authority and the balance
of power, the nature of the mountainous or desert terrain make it feasible. Such
tribalism is politically marginal. It knows what it rejects.
In the Maghreb, as in Zomia, the distinction between a zone of state rule
and a marginal, autonomous zone was geographical and ecological as well as
political. There is a "rough tie-up between high-ground, Berber speech and
political dissidence," such that "gorges and mountains were a clear dividing
line between the land of the government (bled el-makhazen) and the land of
dissidence (bled-es-siba). "64
The Berber case is instructive for two reasons. First, Gellner makes it
abundantly clear that the demarcation line between Arab and Berber is not,
essentially, one of civilization, let alone religion. Instead, it is a political line
distinguishing the subjects of a state from those outside its control. Assum-
ing, as Gellner does, historical movement back and forth across this divide,
what becomes intriguing is that a distinction in political status is ethnically
coded as if it were a fundamental difference in kinds of people and not a po-
litical choice. It means that all those who had reason to flee state power, for
whatever reason, were, in a sense, tribalizing themselves. Ethnicity and tribe
began, by definition, where sovereignty and taxes ended. The ethnic zone
'ras feared and stigmatized by state rhetoric precisely because it was beyond
its grasp and therefore an example ofdefiance and an ever-present temptation
to those who might wish to evade the state.
Gellner's analysis of Berber-Arab -relations is also noteworthy as a long
overdue corrective to what might be called "the view from the valley" or
"the view from the state· center." On that view the "barbarian periphery"
is a diminishing remnant, drawn sooner or later and at varying speeds into
the light of Arab civilization. In Southeast Asia and the Maghreb this view
gains credibility because, in the past century, the ungoverned periphery has
increasingly been occupied by the modern nation-state. Up until then, how-
HILLS, VALLEYS, AND STATES 31
e~er, .the view fro~ the valley-the idea of a luminous and 'magnetic center
ahgnmg and drawmg in peripheral peoples like so many iron filings-is at
the very least, half wrong. Up until then a life outside the state was b~th
more available and more attractive. Oscillation rather than one-way traffic
was the rule. If the account elaborated here emphasizes state avoidance it
is not because that is the whole truth. Rather, it is the largely untold st~ry
that has unfortunately had no legitimate place in the hegemonic narrative of
civilization, despite its historical importance.
This m??el.of s~mbiosis and opposition, of political choice and geo-
graphIcal facIhtatlOn, IS, roughly speaking, applicable to the historical rela-
tionship between hill peoples and valley states in mainland Southeast Asia.
In Southeast Asia, as in the Maghreb, the distinction between the "governed"
~nd the "?n~ove~n~d" is, an apparent social fact, but it is even more firmly
mst~lled m lIngUIstIc usage and popular consciousness. Depending on the
partIcular cultural context, the connotations of the pairs "cooked" and "ravv."
"tame" and" WI, 'ld"" va11ey peop1" e and'" h1l1 people" carry the same weight'
~s makhazen and siba-that is to say, "governed~' and "ungoverned.' The
lmkage between being civilized and being a subject of the state is so taken for
granted that the terms subject peoples on the one hand or self-governing peoples
on the other capture the essential difference.
T?e classical states o~ ~outheast Asia were, as in the Middle East, ringed
by relatIvely free commumtIes: by nonstate spaces and peoples. Such autono-
mous peoples lived not on!y i~ the hills but also in the marshes, swamps,
~angrove co~sts, and labyrmthme waterways of estuarial regions. This mar-
gma! populatlOn represented, at one and the same time, an indispensable
tradmg.partner ~f valley kingdoms, a zone of refuge from state power, a zone
of relatIVe equahty and physical mobility, a source of slaves and subjects for
valley st~tes, ~~d an ecocuIt.ural identity that was nearly a mirror image of
lowl~nd IdentItIes. Thus, whl1e our attention here is trained on the uplands of
ZomIa, we are, more generally, concerned with the relationship between state
spaces and extrastate spaces. The focus on Zomia as a vast interstate mas-
sif, in particular, arises simply because of its importance as the most signifi-
cant comple~ catc~ment zon~ for refugees from state-making projects in the
valley~. ~he mhabitants of thIS zone have come, or remained, here largely be-
cause It hes beyond the reach of the state. Here, the geographical expression
Southeast Asia, as conventionally understood as stopping at the borders of
Southeast Asian nations, is again an impediment to our understanding. Over
the past two millennia, Zomia has been peopled by countless migrations of
32 HILLS, VALLEYS, AND STATES
populations from well beyond its borders - many of them onetime sed.entary
cultivators. They have fled west and southward from Han, a,nd occasIOnally
Tibetan, rule (the Tai, the Yao/Mien, the Hmong/Miao, the Lahu, and the
Akha/Hani) or northward from Thai and Burman rule. Their geographic
location is a political, cultural, and, often, military decision.
I argue further that hill peoples cannot be understood in isolation, say,
as tribes, but only relationally and positionally vis-a-vis valley kingdoms.
Ethnic distinctions and identity in the hills are not only quite variable over
time but' also usually encode a group's relative position vis-a-vis state au-
thority. There are, I would hazard, hardly "tribes" at all, except in this lim-
ited relational sense of the word. The subsistence practices, the choice of
crops to grow, are, by the same token, selected largely with an eye to how,
they facilitate or thwart state appropriation. Finally, as noted earlier, even the
social structures and re~idence patterns in the hills may be usefully viewed
as political choices vis-a-vis state power. Certain egalitarian social structures
reflect, I believe, a Southeast Asian variant of Berber practice: "Divide that
ye be not ruled. "65 Far from being sociological and cultural givens, lineage
practices, genealogical reckoning, local leadership patterns, household struc-
tures, and perhaps even the degrees of literacy have been calibrated to pre-
vent (and in rare cases to facilitate) incorporation in the state.66 A bold case
along these lines is subject to many qualifications and exceptions. I venture
it, nevertheless, not simply to be provocative but because it seems so much
more in keeping with the evidence than the older traditions of relatively self-
contained hill tribes left behind by civilization and progress.
Toward an Anarchist History of Mainland Southeast Asia
What blocks a clear view of the peoples of mainland Southeast Asia for most
of their history is the state: classical, colonial, and independent. While a
state-centric view of, say, the past fifty years might be justified, it represents
a gross distortion of earlier periods. The earlier the period, the greater the
distortion. For most of its history, Southeast Asia has been marked by the
relative absence even of valley states. Where they arose, they tended to be re-
markably short-lived, comparatively weak outside a small and variable radius
of the court center, and generally unable systematically to extract resources
(including manpower) from a substantial population. Indeed, interregna, far
from being uncommon, were more protracted than regna, and, before the
colonial period, a welter of petty principalities allowed much of the popula-
HILLS, VALLEYS, AND STATES 33
tion to shift their residences and loyalties to their advantage or to move to a
zone of no sovereignty or of mutually canceling sovereignties.
Where and when they did exist, the states of mainland Southeast Asia
lurched from solicitous measures designed, to attract subjects to those de-
signed to capture them and ?extract as much grain and labor as possible.
Manpower was the key. Even in those cases wher~ the bulk of the crown's
revenue derived from trade, that revenue was ultimately dependent on the
state's ability to mobilize the manpower to hold and defend an advantageous
position along trade routes. 67 The state was tyrannical, but episodically so.
Physical flight, the bedrock of popular freedom, was the principal check on
state power. As we shall see in some detail, subjects who were sorely tried by
conscription, forced labor, and taxes would typically move away to the hills or
to a neighporing kingdom rather than revolt. Given the vagaries of war, suc-
cession struggles, crop failures, and monarchical delusions of grandeur, such
crises of state-building were unpredictable but, sooner or later, inevitable.
Earlier debates over the writing of Southeast Asian history were about
how the history of states should be written-not about whether states should
have been the center of attention in the first place. Thus scholars criticized
Georges Coedes's Indianized States ofSoutheast Asia for missing the purpose-
ful importation and adaptation of Indian cosmology in the court centers of
Southeast Asia. 68 To the distortions of Indian-centric histories were added,
later, Eurocentric colonial histories in which the local societies were observed
from "the deck of a ship, the ramparts of the fortress, the high gallery of the
trading house. "69 The call was subsequently issued for an "autonomous" his-
tory of Southeast Asia that might avoid both distortions. 70 And yet until very
recently indeed, virtually all the responses to that call have themselves been
histories, however learned and original, of the Southeast Asian state.
Why this should be so, why the histories of states should have so persis-
tently insinuated themselves in the place that might have been occupied by a
history of peoples, merits reflection. The reason, in a nutshell, I believe, is ~hat
state centers, even the tenuous and evan.escent Indic-style classical states, are
the political units that leave the most concentrated volume of physical evi-
dence. The same is the case for sedentary agricultural settlements, character-
istic of state centers. While they are not necessarily any more complex than
foraging or swiddening societies, they are far denser - in the case of irrigated
rice, one hundred times denser - than foraging societies, and hence they
leave far more concentrated rubble in the form of middens, artifacts, build-
ing materials, and architectural ruins.71 The larger the pile of rubble you leave

34 HILLS, VALLEYS, AND STATES
behind, the larger your place in the histori~al rec~r~l T~e more dis.persed,
mobile, egalitarian societies regardless of theIr SOphIStlCa~lOn a~d ~r~dm~ net-
works and despite being often more populous, are relatIvely mVIsible m the
historical record because they spread their debris more widely.72
The same logic applies with a vengeance once it comes to the writte.n
record. Much of what we know about the classical states of Southeast ASIa
comes from the stone inscriptions and, later, paper trails they left behind in
the form of land grants, memorials, tax and corvee reco.rds, religious d~na­
tions and court chronicles.73 The thicker the paper traIl you leave behmd,
the l~rger your place in the historical record. With the written rec~rd, t~e
distortions also multiply. The traditional words in Burmese and ThaI for hIS-
tory,yazawin and phonesavadan, respectively, both.literally. mea~ "the history
of rulers" or "chronicle of kings." It becomes dIfficult, In thIS context, to
reconstruct the life-world of nonelites, even if they are located at the court
center. They typically appear in the record as statistical abstractions: so many
laborers, so many conscripts, taxpayers, padi planters, so many bear~rs of
tribute. Rarely do they appear as historical actors, and when they do, as mthe
case of a suppressed revolt, you can be sure that something has gon~ terribly
wrong. The job of peasants, you might say, is to stay out .of th~ ~rc~Ives.
. Hegemonic histories centered on courts and capItal cItIes mtroduce
other distortions as well. They are, forcibly, histories of "state spaces"; ~hey
neglect or ignore altogether both "nonstate spaces" beyon~ their reach and
the long periods of dynastic decline or collapse when there IS hardly a state at
all. In a truly evenhanded, year-by-year, chronology of precolonial, mainland
Southeast Asian states, most of the pages would be blank. Are we to pre- . project its power only within a fairly small radius ofthe court, say, three hun-
tend, along with the official chronicles, that because there was no dynasty m
control there was no history? Beyond the problem of blank pages, however,
the nat~re of the official histories of the court center systematically exagger-
ates the power, the coherence, and the majesty of the dynasty.74 The court
documents that survive are largely tax and land records on the one hand
and hymns of praise, assertions of power, and claims to legitimacy on the
other; the latter are meant to persuade and to amplify power, not to report
facts. 75 If we take the cosmological bluster emanating from the court centers
as indicative of facts on the ground, we risk, as Richard O'Connor has noted,
"impos[ing] the imperial imaginings of a few great courts on the rest of the
region. "76 . What if we replaced these "imperial imaginings" with a view of South-
The independent nations of mainland Southeast Asia add a new layer
of historical mystification. As the successor states, ethnically and geographi-
HILLS, VALLEYS, AND STATES 35
cally, to the classical,kingdoms, they have their own interest in embellishing
the glory, continuity, and beneficence of their ancestors. Furthermore, the
histories of the classical states have been mined and distorted in the inter-
est of identifying a protonation and a protonationalism that could be of use
against contemporary enemies, both foreign and domestic. Thus early arti-
facts such as Dong Son drums (large bronze ceremonial objects dating from
roughly 500 BeE to the beginning of the common era 'and found throughout
highland Southeast Asia and southern China) or local uprisings have been
appropriated as n~tional and/or ethnic achievements when, at the time, such
identities made no sense at all. The result is an historical fable that projects
the nation and its dominant people backward, obscuring discontinuity, con-
tingency, and fluid identities.77 Such accounts serve, as Walter Benjamin re-
minded us, to naturalize the progre$sion and necessity of the state in general
and the nation-state in particular.78
The inadequacies of mandala, dynastic, capital-city, text-based histo-
ries are so manifest, even when read skeptically, that they are chiefly useful
as self-interested descriptions and cosmological claims. During the greater
part of the historical record, and especially in the uplands, there was no state
or "hardly-a-state." What states there were tended to be personal creations
that were tenuous and fragmented, and that seldom outlasted their founder
by long. Their cosmological claims and ideological reach were far greater than
their practical control over human labor and grain.79
Here it is crucial to distinguish the "hard" power of the state from its
economic and symbolic influence, which was far wider. The precolonial state,
when it came to extracting grain and labor from subject pop~lations, could
dred kilometers, and that undependably and only during the dry season. The
economic reach of the precolonial state, on the other hand, was far wider but
based on voluntary exchange. The higher the value and smaller the weight
and volume of the commodity (think silk and precious gems as opposed to
charcoal or grain), the greater the reach. The symbolic reach of the s'tate -its
regalia, titles, costumes, its cosmology-traveled far and wide as ideas that
have left a deep impression in the hills, even as they were often deployed in
revolts against valley kingdoms. While the valley kingdom's hard power was
a minute fraction of its expansive imperial imaginings, its reach as a market
of physical or, especially, symbolic commodities was far greater.
east Asian history as dominated by long periods of normative and normal-

36 HILLS, VALLEYS, AND STATES
ized statelessness, punctuated by occasional, and usually brief, dynastic states
that, when they dissolved, left in their wake a new deposit of imperial imag-
inings? In a critique of overly state-centric histories, Anthony Day points us
in just this direction: "What would the history of Southeast Asia look like,
however, if we were to take the turbulent relations between families as nor-
mative rather than a departure from the norm of the absolutist state which
must 'deal with disorder'?"80
The Elementary Units of Political Order
Abandoning the tunnel vision of the court-state view, as urged by Day and
O'Connor and actually pursued some considerable distance by Keith Taylor,
we attempt an account of the elementary units of political order in mainland
Southeast Asia. 81 I emphasize the term political order to avoid conveying the
mistaken impression that outside the realm of the state lay mere disorder. De-
pending on the location and date, such units might range from nuclear fami-
lies to segmentary lineages, bilateral kindreds, hamlets, larger villages, towns
and their immediate hinterlands, and confederations of such towns. Confed-
erations appear to constitute the most complex level of integration that had
any stability at all. They consisted of small towns located on terrain favorable
to wet-rice cultivation, with its concentration of population, together with an
allied population in the adjacent hills. Alliances of such "wet-rice archipela-
goes" were common, although they too were short-lived and their constituent
members rarely surrendered their freedom of action. Traces of these pat-
terns survive in place names throughout the entire region: Xishuang Banna
("twelve village rice fields") in Yunnan, Sipsong Chutai ("twelve Tai lords")
along the Vietnamese-Laotian border and Negri Sembilan ("nine realms")
in western Malaysia, and Ko Myo ("nine towns") in Burma's Shan states. In
this respect, the largest quasi-permanent building blocks in the region were
the Malay negeri/Negara, the Tai muang, and the Burmese main (~6:), each
of which represented a potential fund of manpower and grain, located, in the
most favorable cases, athwart a valuable trade route;
Assembling such potential nodes of power into a political and mili-
tary alliance was itself a small, and usually evanescent, miracle of statecraft.
Bringing many such units together under central rule was exceptionally rare
and normally short-lived. When the political confection it represented dis-
integrated, it tended to fragment into its constituent units: the petty state-
lets, small villages, hamlets, and lineages. New agglomerations might arise,
HILLS, VALLEYS, AND STATES 37
orchestrated by a new and ambitious political entrepreneur, but they were
always a contingent alliance of the same elementary units. The symbolic and
ideological format for state-making was known and observed by ambitious
local leaders with even the slightest pretense to wider power. State mimicry-
what I have called cosmological bluster-was copied from the Chinese or
Indic high forms, with rudimentary materials and in miniature, right down
to the most petty village chiefs.
If larger political units were radically unstable, the elementary units
themselves were hardly timeless blocks of building material. We must see
these units themselves as in almost constant motion: dissolving, splitting, re-
locating, merging, and reconstituting. The households and individuals within
a hamlet or lineage were themselves in motion over time. A settlement might
remain in place over, say, half a century, but because of residents coming and
leaving, their linguistic and ethnic identification might shift dramatically.82
Here demography played a central role, the population density in Southeast
Asia being, in 1600, one-sixth that of India and one-seventh that of China.
The existence of an open frontier operated like an automatic brake on what
the state could extract. Motivated by factors as disp'arate as epidemics, fam-
ines, taxes, corvee labor, conscription, factional conflict, religious schism,
shame, scandal, and the desire to change one's luck, it was relatively simple
for households and entire villages to move. Thus, over time, the membership
of any elementary unit was in flux, as was the very existence of the unit itself.
If there was an element of stability here, it resided in the ecology and geogra-
phy of places favorable to human settlements. A well-watered plain situated
on a navigable river or a trade route might occasionally be abandoned, but it
was just as likely to be reinhabited when conditions permitted. Such loca-
tions were, of course, the typical cores of the negeri, the muang, the main.
Fluid as they were, these elementary units were the only building
blocks available to the would-be state-maker. In the absence of an ambitious
strongman, or when the wider polity inevitably shattered, the "remains" , were
once again the elementary units. Is an intelligible history possible under such
circumstances? I believe that it is, although it is surely not a dynastic history.
The units in question do have a history, do observe a rough logic in formation,
combination, and dissolution, and do exhibit a certain autonomy vis-a.-vis dy-
nastic or modern states. They have a history, but that history is on a different
plane from state or dynastic history. For all their fluidity, they are the rela-
tively constant features of the landscape, while'the successful dynastic state is
rare and ephemeral. The contingency of the "state" invites us to treat it less
 HILLS, VALLEYS, AND STATES 39
38 HILLS, VALLEYS, AND STATES

ephemeral, whose subsistence patterns are pliable and fugitive, who have few
as a unity than as a "complex web of contractual mutualities."83 For when
permanent allegiances, and who are liable, over time, to shift their linguistic
it does splinter, as Akin Rabibhadana observed about the early nineteenth- practices and their ethnic identity.
century Siamese state, "the component parts of the system tended to split off
And this is just the point! The economic, political, and cultural orga-
in order to save their own lives."84
nization of such people is, in 'large part, a strategic adaptation to avoid in-
Making sense of innumerable small units, seemingly in constant move-
corporation in state structures. These adaptations are all the more feasible
ment, might seem impossible. It is surely more daunting than dynastic his-
in the mountainous hinterlands of state systems: that,: is to say, in places like
tory, but we are not without guidance from those who have sought to under- Zomia.
stand comparable systems. In the case of Southeast Asia, there are many
studies of social structure that seek to grasp the logic behind the fluidity. Here [Sumatra] I am the advocate of despotism. The strong arm of power is
First, most famous, and most controversial among them is Edmund Leach's necessary to bring men together, and to concentrate them into societies. . ..
Political Systems of Highland Burma. Subsequent work along these lines in Sumatra is, in great measure, peopled by innumerable petty tribes, subject to
the highlands, not to mention studies of the Malay world, where shifting _ no general government.... At present people are as wandering in their habits
petty states, a mobile population, and a distinction between upstream and as the birds of the air, and until they are congregated and organized under
some~hing like authority, nothing can be done with them. ss
downstream, unruled and ruled populations also is at work, is richly sugges-
tive. Beyond Southeast Asia, however, we may look again to the encounter In the early nineteenth century, as in the classical mainland states, Sir Stam-
between states and nomadic, stateless populations in the Middle East. The ford Raffles, quoted above, understood that the precondition of colonial rule
case for beginning with the elementary unit of the household and treating was the concentration of population and sedentary agriculture. He required
villages, tribes, and confederations as provisional and shaky alliances has also a nonfugitive people whose labor and production were legible and hence ap-
been used to brilliant effect for eighteenth-century North American society propriable by the state. We turn our attention next, then, to an understanding
in the Great Lakes region by Richard White. 85 And, finally, we may profita- of the logic and dynamics behind the creation of state spaces in mainland
bly look back to Thucydides' Peloponnesian War, which describes a world of Southeast Asia.
peoples, some with kings, some without, whose fickle loyalties and unreliable
cohesion is a source of constant anxiety to the statesmen of each of the major
antagonists: Athens, Sparta, Corinth, and Syracuse-each of them, in turn,
a confederation.86
One challenge for a non-state-centric history of mainland Southeast
Asia consists in specifying the conditions for the aggregation and disaggre-
gation of its elementary units. The problem has been succinctly put by one
observer ofa somewhat comparable flux between states and their autonomous
hinterlands: "There comes a time when one realizes that one is dealing, really,
with molecules which sometimes unify in the form of a vague confederation,
sometimes, just as easily, disaggregate. Even their names offer no consistency
or certainty."87 Ifthe fluidity ofthe molecules themselves is an inconvenience
for anthropologists and historians, imagine the problem it poses for the dy-
nastic official or would-be state-builder, the colonial official, and the modern
state functionary. State rulers find it well nigh impossible to install an.effec-
tive sovereignty over people who are constantly in motion, who have no per-
manent pattern of organization, no permanent address, whose leadership is

CHAPTER 2
State Space
Zones of Governance and Appropriation
The Geography of State Space and the
Friction of Terrain
Put vegetables in the basket.
Put people in the muang.
- Thai proverb
magine, for a moment, that you are a Southeast Asian counterpart of
I Jean-Baptiste Colbert, chief minister to Louis XlV. You, like Colbert,
are charged with designing the prospe:ity of the kingdom. The settin~,
like that of the seventeenth century, IS premodern: overland travel IS
by foot, cart, and draft animals, while water transportation is by sail. Let us
finally imagine that, unlike Colbert, you begin with a blank slate. You are
free to conjure up an ecology, a demography, and a geography that would be
most favorable to the state and its ruler. What, in those circumstances, would
you design?
Your task, crudely put, is to devise an ideal "state space": that is to say,
an ideal space of appropriation. Insofar as the state depends on taxes or rents
in the largest possible sense of the term (foodstuffs, corvee labor, soldiers,
tribute, tradable goods, specie), the question becomes: what arrangements
are most likely to guarantee the ruler a substantial and reliable surplus of
manpower and grain at least cost?
The principle of design must obviously hinge on the geographical con-
centration of the kingdom's subjects and the fields they cultivate wit~in easy
STATE SPACE 41
reach of the state core. Such concentration is all the more imperative in pre-
modern settings where the economics of oxcart or horse-cart travel set sharp
limits to the distance over which it makes sense to ship grain. A team of
oxen, for example, will have eaten the equivalent of the cartload of grain they
are pulling before they have traveled 250 kilometers over flat terrain. The
logic, albeit with different limits, is captured in an ancient Han proverb: "Do
not make a grain sale over a thousand li'~ -415 kilometers.! The non-grain-
producing elite~; artisans, and specialists at the state's core must, then, be fed
by cultivators who are relatively n~ar. The concentration of manpower in the
Southeast Asian context is, in turn, particularly imperative, and particularly
difficult, given the histo,~icallow population-to-Iand ratio that favors demo-
graphic dispersal. Thus the kingdom's core and its ruler must be defended
and maintained, as well as fed, by a labor supply that is assembled relatively
close at hand.
From the perspective of our hypothetical Colbert, wet-rice (padi,
sawah) cultivation provides the ultimate in state-space crops. Although wet-
rice cultivation may offer a lower r~te of return to labor than other subsis-
tence techniques, its return per unit of land is superior to almost any other
Old World crop. Wet rice thus maximizes the food supply within easy reach
of the state core. The durability and relatively reliable yields of wet-rice cul-
tivation would also recommend it to our Colbert. Inasmuch as most of the
nutrients are brought to the field by the water from perennial streams or by
the silt in the case of "flood-retreat agriculture," the same fields are likely
to remain productive for long periods. Finally, and precisely because wet
rice fosters concentrated, labor-intensive production, it requires a density of
population that is, itself, a key resource for state-making.2
Virtually everywhere, wet rice, along with the other major grains, is the
foundation ofearly state-making. Its appeal to a hypothetical Colbert does not
end with the density of population and foodstuffs it makes possible. From a
tax collector's perspective, grains have decisive advantages over, for examI3!,e,
root crops. Grain, after all, grows aboveground, and it typically and predict-
ably all ripens at roughly the same time. The tax collector can survey the crop
in the field as it ripens and can calcula;te in advance the probable yield. Most
important of all, if the army and/or the'tax collector arrive on the scene when
the crop is ripe, they can confiscate as much of the crop as they wish. 3 Grain,
then, as compared with root crops, is both legible to the state and relatively
. appropriable. Compared to other foodstuffs, grain is also relatively easy to
transpo~t, has a fairly high value per unit of weight and volume, and stores
 42 STATE SPACE
for relatively long periods with less spoilage, especially if it is left unhusked.
Compare, for example, the relative value and perishability of a cartload of
padi, on the one hand, and a cartload of, say, potatoes, cassava, mangoes, or
green vegetables. If Colbert were called on to design, from scratch, an ideal
state crop, he could hardly do much better than irrigated rice.4
No wonder, then, that virtually all of the premodern state cores in
Southeast Asia are to be found in ecological settings that were favorable to
irrigated rice cultivation. The more favorable and extensive the setting, the
more likely a state of some size and durability would arise there. States, it
should be emphasized, did not typically, at least until the colonial era, con-
struct these expanses of padi' fields, nor did they play the major role in their
maintenance. All the evidence points to the piecemeal elaboration of padi
lands by kinship units and hamlets that built and extended the small diver-
sion dams, sluices, and channels required for water control. Such irrigation
works often predated the creation of state cores and, just as frequently, sur-
vived the collapse of many a state that had taken temporary advantage of its
concentrated manpower and food supply.5 The state might batten itself onto
a wet-rice core and even extend it, but rarely did the state create it. The re-
lationship between states and wet-rice cultivation was one of elective affinity,
not one of cause and effect.
The realpolitik behind this elective affinity is evident in the fact that
"for l;:uropean governors and Southeast Asian rulers alike, large settled
,P2pulations supported by abundant amounts of food were seen as the "key
to authority and power."6 Land grants in ninth- and tenth-century Java, for
which we have inscriptional evidence, were made on the understanding that
the recipients would clear the forest and convert shifting, swidden plots into
permanent irrigated rice fields (sawah). The logic, as Jan Wisseman Christie
notes, is that "sawah ... had the effect of anchoring populations and increas-
."" ,ing their visibility, and making the size of the crop relatively stable and easy
to calculate."? No effort was spared, as we shall see in more detail, to attract
and hold a population in the vicinity of the court and to require it to plant
padi fields. Thus Burmese royal edicts of 1598 and 1643, respectively, ordered
that each soldier remain in his habitual place of residence, near the court
center, and required all palace guards not on duty to cultivate their fields. 8
The constant injunctions against moving or leaving the fields fallow are, if
we read such edicts "against the grain," evidence that achieving these goals
met with a good deal of resistance. When such goals were approximated,
however, the result was an impressive "treasury" of manpower and grain at ,
STATE SPACE 43
~e mon~rch's disposal. Such seems to have been the case at Mataram, Java,
m the mid-seventeenth century, when a Dutch envoy remarked on "the un-
believ~bly grea: rice fields which are all around Mataram for a day's travel
and Wlth them mn~me.rable villages." The resources ofmanpower at the cor~
wer.e 110t ~:>nly crucl~l for food production; they were militarily essential to the
defense .and expanslOn. of ~he stat~ against its rivals. The decisive advantage
of agranan state~ of this kl1~d agamst their maritime competitors appears to
have rested ~re~lsely on their numerical superiority in fielding soldiers.
!he fnctlOn of terrain set up sharp, relatively inflexible limits to the
effectlve reach of t~e traditiona~ agrarian state. Such limits were essentially
fixe~, as noted earlier, by the difficulty of transporting bulk foodstuffs. As-
summg level terr~in and good roads, the effective state space would have
become tenuou~ mdeed beyond a radius of three hundred kilometers. In
one ~ense, the difficulty of moving grain long distances, compared with the
relative ease ?f human pedestrian travel, captures the essential dilemma of
S?uthea~t ASian statec~aft b~fore t?e late nineteenth century. Provisioning
t~e state score populatlOn With gram ran up against the intractable limits of
dlstanc~ and har:est fluctuations, while the population sequestered to plant
that gram found it all too easy to walk beyond the reach of state control. Put
another way, the friction and inefficiencies of the oxcart worked to constrict
the food sup~ly ava~lable to the state core, whereas the relatively frictionless
mo:ement ofItS subjects by foot-a movement the premodern state could not
easl1y prevent-threatened to deprive it of grain growers and defenders. 9
!~e stark s~atistical facts of premodern travel and transportation make
the fnctlOn-of-dlstance comparisons between water and land abundanti
~lear. As a rule of thumb, most estimates of travel by foot,assuming an obli;
mgly flat, ?ry terrain, converge around an average of twenty-four kilometers
(fifteen ml1es) ~ day. A strong porter carrying a thirty-six-kilograrn (eighty-
pound) l?ad mIght move nearly as far under very favorable conditions. Once
the terram b.ecom~s ~o:e rugged .orthe weather more challenging (or poth),
h?wever, th1.S op~lm1st1c figure IS dramatically reduced. The ca'lculus is
shghtly modlfied m premodern Southeast Asia, and particularly in warfare
by t~e use of e~ephants, which could carry baggage and negotiate difficul~
terra1~, but theIr numbers were modest and no military campaign depended
essentIally on them. lO
. What might be called state travel through difficult hilly terrain was
conSIderably ~lower. O~e o~ the rare surviving documents (860 CE) from the
Tang dynasty s expanSIOn Into the mountainous areas of mainland South-
 ... )
44 STATE SPACE
east Asia begins with the critical military information about travel times,
expressed in day-stages, between population centers that were nodes of im-
perial control. ll A millennium later, the same preoccupation is apparent. A
representative example is the trip made by Lieutenant C. Ainslie in January
(the dry season) 1892 through the eastern Shan states to assess the political
loyalties of the chiefs and to survey routes of march. He was accompanied
by one hundred military policemen, five Europeans, and a large number of
pack mules, together with their drivers. He used no wheeled transport, pre-
sumably because the tracks were too narrow. Ainslie prospected two parallel
routes between Pan Yang and Mon Pan, a nine-day trip. He reported on
the difficulty of each day's stage and the number of rivers and streams that
had to be crossed, noting in passing that the route was "impassible in the
rains."12 The daily average distance covered was barely more than thirteen
kilometers (eight miles), with considerable daily variation: a maximum ofless
than twenty kilometers and a minimum of barely seven.
A bullock cart can, of course, carry anywhere from seven to ten times
(24 0 -3 60 kilograms) the load of a fit individual porter.13 Its movements, how-
ever, are both slower and more restricted. Where the porter requires only a
footpath, the bullock cart requires a broader track. In some terrain, this is
impossible; anyone familiar with the deeply rutted cart tracks in backcountry
Burma will appreciate how slow and laborious the going is even when such
travel is possible. For a trip of any length the carter must either carry his .own
fodder, thereby reducing the payload, or adjust the route to take advantage
of fodder growing along it.14 Until a century or two ago, even in the West,
the overland transport of bulk commodities "has been subject to narrow and
essentially inflexible limits."15
These geographical givens of movement of people and goods set limits
to the reach of any landward state. Extrapolating from a more generous
estimate of 32 kilometers a day by foot, F. K. Lehman estimates that the
precolonial state's maximum size could not have been much more than 160
kilometers in diameter, although Mataram in Java was considerably broader.
Assuming a court roughly in the center of a circular kingdom with a diame-
ter of, say, 240 kilometers, the distance to the kingdom's edge would be 120
kilometers.16 Much beyond this point, even in flat terrain, state power would
fade, giving way to the sway of another kingdom or to local strongmen andlor
bandit gangs. (See map 3 for an illustration of the effect of terrain on effective
distances.)
Water transport, however, is the great premodern exception to these
STATE SPACE 45
limits. Navigable water nullifies much of the friction of distance. Wind and
currents make it possible to move bulk goods in large quantities over dis-
tances that are inconceivable using carts. In thirteenth-century Europe, ac-
cording to one calculation, shipping costs by sea were a mere 5 percent of the
cost by land. The disparity was so massive as to confer a large strategic and
trade advantage on any kingdom near a navigable water~ay. Most Southeast
Asian precolonial states ofany appreciable size had easy access to the sea or to
a navigable river. In fact, as Anthony Reid notes, the capitals of most South-
I east Asian states were located at river junctions where oceangoing ships had
to transfer their cargoes to smaller craft plying the upstream reaches of the
.river. The location of nodes of power coincided largely with the intersecting
nodes of communication and transportationP
The key role ofwater transportation before the construction of railroads
is evident in the great economic significance of canals, where the draft power
was often the same - horses, mules, oxen - but the reduction of friction made
possible by barges moving over water allowed for huge gains in efficiency.
River or sea transportation takes advantage of "routes of least friction," of
least geographical resistance, and thereby vastly extends the distances over
which food supplies, salt, arms, and people can be exchanged. In epigram-
matic form, we could say that "easy" water "joins," whereas "hard" hills,
swamps, and mountains "divide." ,
Before the distance-demolishing technology ofrailroads and all-weather
motor roads, land-bound polities in Southeast Asia and Europe found it
extremely difficult, without navigable waterways, to concentrate and then
project power. As Charles Tilly has noted, "Before the later nineteenth cen-
tury, land transport was so expensive everywhere in Europe that no country
could afford to supply a large army or big city with grain and other heavy
goods without having efficient water transport. Rulers fed major inland cities
such as Berlin and Madrid only at great effort and great cost to their hinter-
lands. The exceptional efficiency of waterways in the Netherlands undoubt-
edly gave the Dutch great advantages at peace and war."18
The daunting military obstacles presented by travel over very rugged
terrain, even in the mid-twentieth century, was never more· evident than
in the conquest of Tibet by the China's People's Liberation Army in 1951.
Tibetan delegates and party representatives who signed the agreement in
Beijing traveled b~ck to Lhasa via, "the quicker route": namely by sea to
Calcutta, then by train and horseback through Sikkim. Travel from Gong-
tok, Sikkim, to Lhasa alone took sixteen days. Within six months the PLA
 STATE SPACE 47

advance force in Lhasa was in danger of starving, and three thousand tons of
rice was dispatched to them, again by ship to Calcutta and thence by mule
over the mountains. Food came as well from Inner Mongolia to the north,
but this required the astounding mobilization of twenty-six thousand camels,
more than half of whom perished or were injured en route. 19
The standard modern maps, in which a kilometer is a kilometer no
matter what the terrain or body of water, are therefore profoundly misleading
in this respect. Settlements that may be three hundred or four hundred kilo-
meters distant over calm, navigable water are far more likely to be linked by
social, economic, and cultural ties than settlements a mere thirty kilometers
away over rugged, mountainous terrain. In the same fashion, a large plain
that is easily traversed is far more likely to form a coherent cultural and social
whole than a small mountainous zone where travel is'slow and difficult.
Were we to require a map that was more indicative of social and eco-
nomic exchange, we would have to devise an entirely different metric for
mapmaking: a metric that corrected for the friction of terrain. Before the mid-
nineteenth century revolution in transportation, this might mean construct-
ing a map in which the standard unit was a day's travel by foot or oxcart (or
by sailing vessel). The result, for those accustomed to standard, as-the-crow-
flies maps, would look like the reflection in a fairground funhouse mirror.2o
Navigable rivers, coastlines, and flat plains would be massively shrunken to
reflect the ease of travel. Difficult-to-traverse mountains, swamps, marshes,
and forests would, by contrast, be massively enlarged to reflect travel times
even though the distances, as the crow flies, might be quite small. Such maps,

Map 3. The striking constriction of state space imposed by rugged landscape may be
illustrated by a map 'that compares wa~king times from a central place, depending on
the difficulty of the terrain. Here we have selected Mung (Muang) Yang, a Shan town
near the Burma-Chinese border, for illustrative purposes. The walking-time isolines
shown here are based on Waldo Tobler's "hiker function," an algorithm that esti111ates
the rate of travel possible based upon the slope at any given point on the landscap'e.
These isolines show the travel distance possible assuming a six-hour walking day. The
travel distances possible on flat terrain, based upon the Tobler algorithm, are shown in
dotted lines for 'comparison. Setting out from Mung Yang, a traveler takes three days
to cover the distance that, were the land flat, one could cover in a day and a half or two
days. Travel is more difficult to the south and northwest than to, the east. If we assume
that the span ofcontrol varies directly with the ease of travel, than the total area under
control of a hypothetical statelet centered on Mung Yang would be less than one-third
of what it might be over level terrain.
48 STATE SPACE
however strange to the modern eye, would be far superior guides to contact,
culture, and exchange than the ones to which we have grown accustomed.
They would also, as we shall see, help demarcate the sharp difference between
a geography more amenable to state control and appropriation (state space)
and a geography intrinsically resistant to state control (nonstate space).
A map in which the unit of measurement is not distance but the time of
travel is, in fact, far more in accord with vernacular practices than the more
abstract, standardized concept of kilometers or miles. If you ask a Southeast
Asian peasant how far it is to the next village, say, the answer will probably be
in units of time, not of linear distance. A peasant quite familiar with watches
might answer "about half an hour," and an older farmer, less familiar with
abstract time units, might reply in vernacular units, "three rice-cookings"
or "two cigarette-smokings" -units of duration known to all, not requiring
a wristwatch. In some older, precolonial maps, the distance between any two
places was measured by the amount of time it took to travel from one to the
other. 21 Intuitively this makes obvious sense. Place A may be only twenty-five
kilometers from place B. But depending on the difficulty of travel, it could be
a two-day trip or a five-day trip, something a traveler would most surely want
to know. In fact, the answer might vary radically depending on whether one
was traveling from A to B or from B to A. If B is in the plains and A is high
in the mountains, the uphill trip from B to A is sure to be longer and more
arduous than the downhill trip from A to B, though the linear distance is the
same.
A friction of distance map allows societies, cultural zones, and ~ven
states that would otherwise be obscured by abstract distance to spring ,sud-
denly into view. Such was the essential insight behind Fernand Braudel's
analysis of The Mediterranean World. Here was a society that maintained itself
by the active exchange of goods, people, and ideas without a unified "terri-
tory" or political administration in the usual sense of the term. 22 On a some-
what smaller scale, Edward Whiting Fox argues that the Aegean of classical
Greece, though never united politically, was a single, social, cultural" and
economic organism, knit together by thick strands ,of contact and exchange
over easy water. The great "trading-and-raiding" maritime peoples, such as
the Viking and Normans, wielded a far-flung influence that depended on fast
water transport. A map of their historical influence would be confined largely
to port towns, estuaries, and coastlines.23 Vast sea spaces between these would
be small.
The most striking historical example ofthis phenomenon was the Malay
world - a seafaring world par excellence- whose cultural influence ran all the
STATE SPACE 49
way from Easter Island in the Pacific to Madagascar and the coast of Southern
Africa, where the Swahili spoken in the coastal ports bears its imprint. The
Malay state itself, in its fifteenth- and sixteenth-century heyday, could fairly
be called, like the Hanseatic League, a shifting coalition of trading ports. The
elementary units of statecraft were ports like Jambi, Palembang, Johor, and
Melaka, and a Malay aristocracy shuffled between them; depending on politi-
cal and trade advantages. Our landlocked sense of a "kingdom" as consisting
of a compact and contiguous territory makes no sense when confronted with
such maritime integration across long distances.
An agrarian kingdom is typically more self-contained than a maritime
kingdom. It disposes of reserv~s of food and manpower close to home. N ever-
theless, even agrarian kingdoms are far from self-sufficient; they depend for
their survival on products outside their direct control: hill and coastal prod-
ucts such as wood, ores, protein, manure from pastoralists' flocks, salt, and so
on. Maritime kingdoms are even more dependent on trade routes to supply
their necessities, including, 'especially, slaves. For this reason, there are what
might be called spaces of high "stateness" that do not depend on local grain
production and manpower. Such locations are strategically situated to facili-
tate the control (by taxes, tolls, or confiscation) of vital trade products. Long
before the invention of agriculture, those societies controlling key deposits
of obsidian (necessary for the best stone tools) occupied a privileged position
in terms of exchange and power. More generally thyre were certain strategic
choke points on land and water trade routes, the control of which might con-
fer decisive economic and political advantages. The Malay trading port is the
classical example, typically lying athwart a river junction or estuary,allowing
its ruler to monopolize trade in upstream (hulu) export products and simi-
larly to control the hinterland's access to trade goods from downstream (hilir)
coastal and international commerce. The Straits of Malacca were, in the same
fashion, a choke point for long-distance trade between the Indian Ocean and
China and thus a uniquely privileged space for state-making. On a smaller
scale, innumerable hill kingdoms sat astride important caravan routes 'for salt,
slaves, and tea, among other goods. They waxed and waned depending on
the vagaries ofworld trade and commodity booms. Like their larger Malay
cousins, they were, at their most peaceful, "toll" states.
Positional advantages of this kind are only partly a matter of ~he terrain
and sea lanes. They are, especially in the modern era, historically contingent
0n revolutions in transport, engineering, and industry: for example, rail and
road junctions, bridges and tunnels, coal, oil, and natural gas deposits.
Our crude first approximation ofstate space as the concentration ofgrain
50 STATE SPACE
production and manpower in a manageable space must then be modified. The
distance-demolishing properties of navigable water routes and the existence
of nodes of power represented by choke points and strategic commodities
can compensate for deficiencies in grain and manpower close at hand, but
only to a point. Without sufficient manpower, it is frequently difficult for toll
states to hold onto the site that confers a positional advantage. In the case of a
showdown, agrarian states have generally been able to prevail over maritime
or "trade-route" states by force of numbers. The disparity is highlighted by
Barbara Andaya's comparison of the Vietnamese Trinh (an agrarian state)
and Johore (a maritime state) at the beginning of the eighteenth century:
"The point can be made clearly by comparing the armed forces of Johore,
the most prestigious of the Malay States, but one without any agrarian base,
with those of the Trinh. In 1714, the Dutch estimated that Johor could bring
into battle 6,500 men and 233 vessels of all types. In Vietnam~ by contrast,
the Nguyen army was tallied at 22,740 men, including 6,400 marines and
3,280 infantry." 24 The earliest cautionary tale of maritime-state vulnerability
is, of course, Thucydides' Peloponnesian War, in which a resolutely maritime
Athens is, finally, undone by its more agrarian rivals, Sparta and Syracuse.
Mapping State Space in Southeast Asia
State-building in precolonial mainland Southeast Asia was powerfully con-
strained by geography. Here, in a rough and ready way, I shall attempt to
outline those major constraints and their effects on the location, maintenance,
'and power dynamics of such states.
The necessary, but by no means sufficient, condition for the rise of a
substantial state was the existence of a large alluvial plain suitable for the
cultivation of irrigated rice and hence capable of sustaining both a substantial
and concentrated population. Unlike maritime peninsular Southeast Asia,
where the ease of movement over the calm waters of the Sunda Shelf per-
mitted the coordination of a far-flung thalassocracy on the order of Athens,
mainland states had to contend with far higher levels of geographical fric-
tion. Because of the generally north-south direction of mountain ranges and
major rivers in the region, .virtually all of the classical states were to be found
along the great north-south river systems. They were, moving from west to
east, the Burman classical states along the Irrawaddy near its confluence with
the Chindwin (Pagan, Ava, Mandalay) or along the Sittang not far to the
east (Pegu, Toungoo); the Thai classical state (Ayutthaya and, much later,
Bangkok, along the Chao Phraya); the Khmer classical state (Angkor and its
STATE SPACE 51
successors) near the great lake of TonIe Sap, a tributary of the Mekong; and
finally, the early heartland of the Kinh (Trinh) classical state along the Red
River in the vicinity of Hanoi.
The common denominator here is that all such states have been created
near navigable water courses, but above the flood plain, where a flat, arable
plain and perennial streams made wet-rice cultivation possible. It is striking
that none of the early mainland states was located in the 'delta of a major river.
Such delta regions-the Irrawaddy, the Chao Phraya, and Mekong-were
settled in force and planted to wet rice only in the early twentieth century.
The reasons for their late development, apparently, are that 1) they required
extensive drainage works to be made suitable for rice cultivation, 2) they
were avoided because they were malarial (especially when newly cleared), and
3) the annual flooding was unpredictable and often devastating. 25 This bold
generalization, however, needs to be clarified and qualified. First, the political,
economic, and cultural influence emanating from such centers of power, as
Braudel would have predicted, spread most easily when least impeded by the
friction of distance-along level terrain and navigable rivers and coastlines.
Nothing illustrates this process more strikingly than the gradual, intermit-
tent displacement of Cham and Khmer populations by the Vietnamese. This
expansion followed the thin coastal strip southward, with the coast serving
as a watery highway leading, eventually, all the way to the Mekong Delta and
the trans-Bassac. -
The economic reach of such state centers was almost always greater
than their political reach. While their political. control was limited by their
degree of monopoly access to mobilized manpower and food supplies, their
influence on trade might reach considerably farther. The friction of distance
is at work here too; the greater the exchange value of a product vis-a-vis its
weight and volume, the greater the distance over which it might be traded.
Thus precious commodities such as gold, gemstones, aromatic woods, rare
medicines, tea, and ceremonial bronze gongs (important prestige goods in
the hills) linked peripheries to centers on the basis of exchange rather than
political domination. On this basis, the geographical scope ofcertain forms of
trade and exchange, requiring no bulk transport, was far more extensive than
the comparatively narrow range within which political integration might be
achieved.
I have thus far considered only the major classical states in mainland
Southeast Asia. The key condition for state formation was present elsewhere
as well: a potential heartland of irrigated rice cultivation that might consti-
tute a "f1:lly-administered territorial nucleus, having a court capital at its
 STATE SPACE 53

center. "26 The difference was purely a matter of scale. Where the heartland
of irrigated rice was large and contiguous, it might, under the right condi-
tions, facilitate the rise of a major state; where the heartland was modest, it
might, also under the right conditions, give rise to a modest state. A state on
this account would be a fortified town of, say, at least six thousand subjects
plus nearby hill allies, situated on wet-rice plain an,d having, in theory at
least, a single ruler. Scattered throughout mainland'Southeast Asia, often
at fairly high altitudes, one finds the agro-ecological conditions that favor
state formation, usually on a more Lilliputian scale. Most such places were
at one time or another the sites of small Tai statelets. More rarely, leagues or
confederacies of such statelets might combine, briefly, to forge a more formi-
dable state. State formation around wet-rice cores, large or small, was always
contingent and, typically, ephemeral. One might emphasize with Edmund
Leach the fact that "the riceland stayed in one place" and thus represented a
potential ecological and demographic strong point, which a clever and lucky
political entrepreneur might exploit to create a new, or revived, state space.
Even a successful dynasty was by no means a Napoleonic state; it was rather
a shaky hierarchy of nested sovereignties. To the degree that it held together,
the glue was a prudent distribution ofspoils and marriage alliances and, when
necessary, punitive expeditions for which, in the final analysis, control over
manpower was vital.
Our conception of what constituted precolonial Burma must therefore
be adjusted according to these basic principles of appropriation and span of
control. Under a robust, flourishing dynasty, "Burma," in the sense of an
effective political entity, consisted largely of wet-rice core areas within a few
days' march from the court center. Such wet-rice areas need not necessarily
be contiguous, but they had to be relatively accessible to officials and sol-
diers from the center via trade routes or navigable waterways. The nature
of the routes of access was itself crucial; an army on its way to collect grain
or to punish a rebellious district had to provision itself en route. This meant

Map 4. Rivers and classical states of Southeast Asia: The coincidence of classical states
with navigable water courses is the general rule, as the map' illustrates. The Salween/
Nu/Thanlwin River spawned only one classical state, Thaton, at its estuary. For much
of its long course, the Salween runs through deep gorges and is not navigable. It is,
solely for this reason, an exception. Keng Tung and Chiang Mai are also exceptions
in the sense that neither is located close to a major navigable river. Each, however,
commands a large, arable plain suitable for padi cultivation and hence for state-making.
54 STATE SPACE

locating a route of march through territory sufficiently rich in grain, draft

animals, carts, and potential recruits for the army to sustain itself. .
Thus marshes, swamps, and, especially, hilly areas, though they might
be quite close to the court center, were generally not a part of "political, di-
rectly administered Burma. "27 Such hills and marshes were sparsely popu-
lated and, except in the case of a substantial plateau suitable for 'irrigated rice,
their population practiced a form of mixed cultivation (dispersed swiddens
for hill rice, root crops, foraging, and hunting) that was difficult to assess,
let alone appropriate. Areas of this kind might have a tributary alliance with
the court specifying the periodic renewal of oaths and the exchange of valu-
able goods, but they remained generally outside the direct political control
of court officials. As a rule of thumb, hilly areas above three hundred meters
in elevation were not a part of "Burma" proper. We must therefore consider
precolonial Burma as a flatland phenomenon, rarely venturing out of its
irrigation-adapted ecological niche. As Braudel and Paul Wheatley noted in
general, political control sweeps readily across a flat terrain. Once it confronts
the friction of distance, abrupt changes in altitude, ruggedness of terrain, and
the political obstacle of population dispersion and mixed cultivation, it runs
out of political breath.
Modern concepts of sovereignty make little sense in this setting. Rather
than being visualized as a sharply delineated, contiguous territory following
the mapmaking conventions for modern states, "Burma" is better seen as a
horizontal slice through the topography, taking in most areas suitable for wet
rice below three hmldred meters and within reach of the court. 28 .
Imagine a map constructed along these lines, designed to represent rela-
tive degrees of potential sovereignty and cultural influence. One way of visu-

Map s. Elevation in central Burma: The "reach" of the precolonial state, at its most
robust, stretched most easily along the low elevation plains and navigable river courses.
All of the upper Burma kingdoms hugged the Irrawaddy above or below its confluence ELEVATION
with the Chindwin. The Shan Hills to the east of Mandalay and Ava, though closer IN METERS
as the crow flies than the downriver towns ofPokokku and Magway, were outside the
effective limits of the kingdom. The precolonial state also skirted the north-south Pegu- 11& 0-200
Yoma range of modest but rugged hills that bisected the rice plain. These hills remained
effectively outside state control in the precolonial period, in much of the colonial
11& 201-5 00

period, and in independent Burma, where they were the redoubt of communist and 501- 1,500
Karen rebels until 1975. It is a striking example of how even relatively modest changes
in the friction of terrain can impede state control.
W 1,501 - 8,752 o 200 KM
 STATE SPACE 57

alizing how the friction of distance might work is to imagine yourself holding
a rigid map on which altitudes were represented by the physical relief of the
map itself. Further, let's imagine that the location of each rice-growing core
is marked by a reservoir of red· paint filled to the very brim. The size of the
reservoir of paint would be proportional to the size of the wet-rice core and
hence the population it might accommodate. Now visualize tilting this map,
now in one direction, now in another, successively. The paint as it spilled
from each reservoir would flow first along level ground and along the lowland
water courses. As you increased the angle at which the map was tilted, the
red paint would flow slowly or abruptly, depending on the steepness of the
terrain, to somewhat higher elevations.
The angle at which you had to tilt the map to reach particular areas
would represent, very roughly, the degree of difficulty the state would face in
trying to extend its control that far. If we assume that the intensity of the red
fades both in proportion to the distance it has traveled and the altitude it has
attained, we have an approximation, again very roughly, of the diminishing
influence and control or, alternatively, the relative cost of establishing direct
political control in such areas. At higher elevations~' the red would give way
to white; if the terrain there was both steep and high, the transition would
be quite abrupt. From above, depending on the number of hilly areas near
. the court center, this depiction of sovereignty would reveal a number of ir-
regular white spots against a dark or pale red background. The population
that inhabited the white blotches, although it might often be in a tributary
relation to the court center, was rarely if ever directly ruled. If political con-

Map 6. Minbu Kharuin (K'a yain) and Kyaukse irrigation works: These two main
irrigation zones were the rice basket of precolonial states in upper Burma. The Minbu
Kharuin irrigation works considerably predate the Pagan kingdom's rise in the ninth
century CEo These two rice cores formed the repository of manpower arid grain
necessary to state formation and its inevitable accompaniment, warfare. (The term
k'd yain-Ql~8-often transliterated kharuin, means "district" and connotes a walled
town, as in the famous "nine K'a yain" making up classical Kyaukse. Ii is the equivalent
in most respects of the Shan term main-~8:-or the Thai muang.) Outside these
two zones, on the plain, there was rain-fed, arable land, but the yields were neither as
reliable nor as bounteous as those from the irrigated lands. In the north salient of the
Pegu Yoma-Mount Popa and the elevated hills extending from it-population and
agricultural production were even sparser. And the population and produce present
were difficult to appropriate.

58 STATE SPACE
trol weakened suddenly before the daunting hills, cultural influence weak-
ened as well. Language, settlement patterns, kinship structure, ethnic self-
identification, and subsistence practices in the hills were distinctly different
from those in the valleys. For the most part, hill peoples did not follow valley
religions. Whereas the valley Burmans and Thais were Theravada Buddhists,
hill peoples were, with some notable exceptions, animist and, in the twentieth
century, Christians.
The color scheme of this fantasy friction-of-distance map would also
offer a rough and ready guide to patterns of cultural and commercial, but not
political, integration. Where the red color spreads with the least resistance,
along river courses and flat plains, there one is .likely to find more homo-
geneity in religious practices, language dialects, and social organization as
well. Abrupt cultural and religious changes are likely to occur at the same
places where there is, as with a mountain range, an abrupt incrt:ase in the fric-
tion of distance. If the map could also show, like a time-lapse photograph, the
volume of human and commercial traffic across a space as well as the relative
ease of movement, we wOilld have an even better proxy for the likelihood of
social and cultural integration. 29
Our metaphorical map, like any map, though it serves to foreground
the relationships we wish to highlight, obscures others. It cannot easily ac-
count, in these terms, for the friction of distance represented, say, by swamps,
marshes, malarial zones, mangrove coasts, and thick vegetation. Another cau-
tion concerns the "pot of paint" at the state core. It is purely hypothetical;
it represents the plausible reach of influence of a vigorous, ambitious state
core under the most favorable conditions. Few state cores even came close to
realizing this degree of sway over their hinterlands.
None of these state cores, large or small, had the terrain to itself. Each
existed as one unit among a galaxy of waxing and waning contending centers.
Before colonial domination and the codification of the modern territorial state
vastly simplified the terrain, the sheer numbers of state centers, mostly Lil-
liputian, was bewildering. Leach was not exaggerating when he noted that
"practically every substantial township in 'Burma' claims a history of havin~
been at one time or another the capital of a 'kingdom' the alleged frontiers of
which are at once· both grandiose and improbable."30
How might we represent, again schematically, this plurality of state
centers? One alternative is to invoke the Sanskritic term mandala ("circle of
kings"), much used in Southeast Asia, in which the influence ofa ruler; often
claiming divine lineage, emanates from a court center, almost always located
STATE SPACE 59
on a rice plain, out into the surrounding countryside. In theory, he rules over
lesser kings and chiefs who recognize his claim to spiritual and temporal au-
thority. The anachronistic metaphor of a light bulb with varying degrees of
illumination to represent the charisma and sway of a ruler, first suggested by
Benedict Anderson, captured two essential features ofmandala-style political
Its
centers.3! dimming suggested the gradual diminution of power, both spiri-
tual and temporal, with distance from the center, and its diffuse glow avoided
any modern assumption of "hard" boundaries within which 100 percent sov-
ereignty prevailed and beyond which it disappeared altogether.
In figure I I attempt to depict some of the striking complexities of sov-
ereignty in a plural mandala system. In order to do so, I have represented
a number of mandala (negara,. muang, main, k'it yain) by fixed circles with
power concentrated at the center and fading gradually to zero at the outer
circumference. This requires us, for the moment, to overlook the massive
influence of terrain. We assume, in effect, a plain as flat as a pancake. Bur-
mese authorities in the seventeenth century also made such simplifying as-
sumptions in their own territorial order: a province was imagined as a circle
and specified to have an administrative radius of exactly one hundred tz'ang
(one tiang equals 3'14 kilometers), a big town a radius often tiang, a medium
town five tiang, and a village two and a half tiang.32 The reader should imag-
ine how geographical irregularities-say, a swamp or rugged terrain-would
truncate these circular shapes or how a navigable river might extend their
reach along the waterway. More egregious, still, the very fixity of the repre-
sentation of space completely ignores the radical temporal instability of the
system: the fact that "centers ofspiritual authority and political power shifted
endlessly."33 The reader should rather imagine these centers as sources of
light thai blaze, go faint, and are in time extinguished altogether, while new
sources of light, points of power, suddenly appear and glow brighter.
Each circle represents a kingdom; some are smaller, others are larger,
but the power of each recedes as one moves to the periphery, as represented
by the diminishing density of icons within each mandala. The purpos·e of this
rather facile graphic is merely to illustrate some of the complexities of power,
territory, and sovereignty in precolonial mainland Southeast Asia, worked
out in considerably more detail by Thongchai WinichakuP4 In theory, the
lands within a mandala's sway provided an annual tribute (which might be re-
ciprocated by a gift ofequal or greater value) and were obliged to send troops,
. carts, draft animals, food, and other supplies when required. And yet, as the
graphic indicates, many areas fell within the ambit ofmore than one overlord.

60 STATE SPACE
I. Schema of mandalas as fields of power
Where dual sovereignty, as in the area D / A, was located at the periphery of
both kingdoms, it might well represent a case of mutually canceling, weak
sovereignty, affording local chiefs and their following great autonomy in this
buffer area. Where it affected much of the kingdom, as in B/A or A/C, it
might be the occasion of competing exactions and/or punitive raids by the
center on noncompliant, disloyal villages. Many hill peoples and pet1:y chief-
taincies strategically manipulated the situation of dual sovereignty, quietly
sending tributary missions to two overlords and representing themselves to
STATE SPACE 61
their own tributaries as independent. 35 Calculations oftribute were not an all-
or-nothing affair, and the endless strategic choices of what to send, when to
send it, when to delay, when to withhold manpower and supplies were at the
very center of this petty statecraft.
Outside the central core of a kingdom, dual or multiple sovereignty or,
especially at higher elevations, no sovereignty, was less an anomaly than the
norm. Thus Chaing Khaeng, a small town near the crtrrent borders of Laos,
Burma, and China, was tributary to Chiang Mai and Nan (in turn, tributary
to Siam) and to Chiang Tung/Keng Tung (in turn, tributary to Burma). The
situation was common enough that small kingdoms were often identified as
"under two lords" or "under three lords" in the Thai language and its Lao
dialect, and a "two-headed bird" in the case of nineteenth-century Cambo-
dia's tributary relationship to both Siam and Dai Nan (Vietnam).36
Unambiguous, unitary sovereignty, of the kind that is normative for the
twentieth-century nation-state, was rare outside a handful of substantial rice-
growing cores, whose states were, themselves, prone to collapse. Beyond such
zones, sovereignty was ambiguous, plural, shifting, and often void altogether.
Cultural, linguistic, and ethnic affiliations were, likewise, ambiguous, plural,
and shifting. If we add to this observation what we understand about the
friction of terrain and altitude in projecting political power, we can begin to
appreciate the degree to which much of the population, and most especially
the hill peoples, were, though never untouched by the court centers of the
region, hardly under their thumb.
Even the most robust kingdom, however, shrank virtually to the ram-
parts of its palace walls once the monsoon rains began in earnest. The South-
east Asian state, in its precolonial mandala form, its colonial guise, and, until
very recently, as a nation-state, was a radically seasonal phenomenon. On the
ma~nland, roughly from May through October, the rains made the roads im-
passable. The traditional period for military campaigns in Burma was from
November to February; it was too hot to fight in March and April, and from
May through much of October it was too rainy.37 Not only were ar'mies and
tax collectors unable to move far in any force, but travel and trade were re-
duced to a trivial proportion of their dry-season volume. To visualize what
this meant, we would have to consider our mandala map as a dry-season rep-
resentation. For the rainy season, we would have to shrink each kingdom to
something like a quarter to an eighth of its size, depending on the terrain. 38
. As if some semiannual flood. tide virtually marooned the state as the rains
began and then released its watery grip when they stopped, state space and
62 STATE SPACE
STATE SPACE 63

nonstate space traded places with meteorological regularity. A hymn of praise mount a punitive ~xpedition, bur?ing houses and aboveground crops, but
to a fourteenth-century Javanese ruler notes the periodicity of rule: "Every long-ter.m occupatIOn was beyond Its reach. Unless it had hill allies a hostile
time at the end of the cold season [when it is quite dry] he sets out to roam populatI?n need only wait for the rains, when supply lines broke 'down (or
through the countryside. . . . He shows the flag especially in remote areas. were eaSIer to .cut) and ~he garrison was faced with starvation or retreat. 42
... He displays the splendor of his court.... He receives homage from all Thus t~e p.hyslcal, coerCIve presence of the state in the remotest, hilly areas
and sundry, collects tribute, visits village elders, checks land registers and was epIsodIC, often to the vanishing point. Such areas represented a reliable
examines public utilities such as ferries, bridges and roads."39 Subjects knew zone of refuge for those who lived there or who chose to go there.
roughly when to expect their ruler. They also knew roughly when to expect
armies, press-gangs, military requisitions, and the destruction of war. War,
like fire, was a dry-season phenomenon. Military campaigns, such as the sev-
eral invasions ofSiam by the Burmese, always began after the end ofthe rainy
season, when the tracks were again passable and the crops ripening. 40 Any
thorough examination of traditional state-making would have to give almost
as large a place to weather as to pure geography.
Colonial regimes, though they worked mightily to construct all-weather
roads and bridges, were thwarted in much the same way as the indigenous
states they replaced. In the long, arduous campaign to occupy upper Burma,
the progress made by colonial troops (mostly from India) in the dry season
was often undone by the rains and, it seems, by the diseases of the wet season
as well. An account of the effort in 1885 to clear Minbu, in upper Burma,
of rebels and bandits, revealed that the rains forced a withdrawal of British
troops: "And by the end of August the whole of the western part ofthe district
was in the hands of the rebels and nothing remained to us but a narrow strip
along the river-bank. The rains and the deadly season which succeeds them
in the water-logged country at the foot of the Yoma [Pegu-Yoma mountain
range] . > • prevented extended operations from being undertaken before the
end of the year [again the dry season]."41 In the steep, mountainous terrain
along the Thai border where the Burmese army today fights a war without
mercy against its ethnic adversaries, the rainy season remains a major handi-
cap to regular armed forces. The typical offensive "window" for Burmese
troops has been exactly that of the former kings ofPagan and Ava: November
through February. Helicopters, forward bases, and new communications gear
have allowed the regime to mount, for the first time, wet-season offensives.
Nevertheless, the capture of the last major Karen base on Burmese territory
took place on January 10, 1995, just as the earlier pattern of seasonal warfare
would have dictated.
For those wishing to keep the state at arm's length, inaccessible moun-
tain redoubts constituted a strategic resource. A determined state might

Apr 13, 2010